Every criminal case in India moves through a sequence of decision points, and almost every one of those decision points is governed by a single statute, the Code of Criminal Procedure 1973. A report reaches a police station and someone must decide whether it becomes a first information report. A suspect is identified and someone must decide whether arrest is necessary or whether a written notice will do. Days pass in custody and a magistrate must decide, at fixed intervals, whether detention continues. An investigation ends and a court must decide whether the material justifies putting a person on trial at all. The Code of Criminal Procedure 1973 is the text that disciplines each of these decisions, attaching a provision, an actor and very often a time limit to every step between the first report of a crime and the final judgment. The substantive offences themselves, the acts the criminal law forbids and the punishments it prescribes, live in a companion statute, the Indian Penal Code, whose complete guide is the substantive half of this pair; the Code of Criminal Procedure supplies the machinery through which those offences are investigated, charged and tried.

The brief’s central image for this statute deserves to be stated at the outset, because it organises everything that follows. The Code is a clock with two hands. One hand measures the investigation against the default bail deadline, the period within which the police must complete their work and file their report or lose the accused to bail by right. The other hand measures the liberty of the accused against the remand cycle, the rhythm of production before a magistrate at which custody must be justified again and again rather than assumed. Almost every contested criminal proceeding in India is, underneath its facts, a fight over which hand runs out first. A prosecution that treats the investigation deadline as elastic discovers that the statute converts its delay into the accused’s release. A defence that ignores the remand cycle discovers that custody, once authorised mechanically, hardens into months of undertrial detention that no later acquittal can give back. Reading the Code as a clock rather than as an index of sections is the difference between knowing its contents and understanding its operation.

One structural achievement stands above the rest, and it is the reform that summaries of this statute most often omit. The Code of Criminal Procedure 1973 separated the judiciary from the executive. Under the Code it replaced, the district magistrate, an executive officer answerable to the state government, stood at the head of criminal justice in the district and magistrates who tried cases were part of an administrative hierarchy the executive controlled. The 1973 Code broke that arrangement. Judicial Magistrates, who conduct trials and decide custody, were placed under the control of the High Court, while Executive Magistrates, who handle preventive and administrative functions, remained under the state government. That division, recommended by the Law Commission and rooted in the directive principle in Article 50 of the Constitution, is the single feature that most distinguishes the 1973 Code from the Code of 1898, and it shapes everything from bail hearings to prohibitory orders. This article follows the Code from the first information report to judgment as that sequence of timed decision points, explains the separation and where it remained incomplete, and traces the statute through its amendments to its replacement by a successor code in 2024, after which it continues to govern the large body of cases that were already pending.

What Existed Before the Code of Criminal Procedure 1973

Criminal procedure in India was codified in stages during the nineteenth century, and the Code of 1973 was the fifth general settlement of the subject rather than the first. A Code of Criminal Procedure was enacted in 1861, replaced in 1872, replaced again in 1882, and consolidated in the Code of Criminal Procedure 1898, which then governed criminal process for three quarters of a century. The 1898 Code was, by the standards of its time, a complete and workable instrument, and much of its architecture survives recognisably in its successor: the division between cognizable and non-cognizable offences, the powers of arrest, the remand machinery, the charge sheet, the distinction between summons cases and warrant cases. What changed between 1898 and 1973 was not the existence of the machinery but the constitutional setting in which it had to operate and the identity of the hands that worked it. A code drafted for a colonial administration, in which the district officer combined executive and judicial authority as a matter of design, had to be rethought for a republic whose Constitution promised an independent judiciary, guaranteed personal liberty in Article 21, and directed the state in Article 50 to separate the judiciary from the executive in the public services.

The deeper defect of the older arrangement was the fused magistracy. In the districts, the magistrates who conducted criminal trials were subordinate to the District Magistrate, and the District Magistrate was the head of the district executive, responsible to the state government for order and administration. A trial magistrate’s career, postings and prospects ran through the very executive whose police had investigated the case and whose prosecution he was trying. The objection was not that every such magistrate decided cases at the government’s bidding; most plainly did not. The objection was structural. Justice must not only be done but must be structured so that its independence does not depend on the character of individuals, and a system in which the trying officer sits inside the prosecuting administration fails that test on its face. Successive committees and commissions before and after independence returned to this point, and the Law Commission of India made it the centre of the most thorough examination the 1898 Code ever received.

That examination was the Law Commission’s Forty-first Report, submitted in 1969, which reviewed the 1898 Code section by section and recommended a comprehensive revision. The Forty-first Report is the drafting basis of the 1973 Code, and two of its deliberate reforms define the statute. The first is the separation of the judiciary from the executive, carried into the new Code’s scheme of Judicial Magistrates under the High Court and Executive Magistrates under the state government. The second is the creation of anticipatory bail, a jurisdiction that had no equivalent in the 1898 Code at all. The Commission recommended it because of an observed practice that disfigured the criminal process: persons were implicated in false cases, sometimes by rivals and sometimes in the course of political or property disputes, precisely in order to secure their arrest and detention and the humiliation that attended it, even where the accusation later collapsed. A remedy that operated only after arrest came too late to answer that abuse, so the Commission proposed that a court be empowered to direct, in advance, that a person who was arrested on such an accusation be released on bail. Parliament accepted both reforms. The Code of Criminal Procedure 1973 received the assent of the President on 25 January 1974 and was enacted as Act 2 of 1974, and it came into force on 1 April 1974, repealing and replacing the Code of 1898 across the whole field of criminal procedure.

The new Code also carried quieter reforms that matter in practice. It strengthened the position of the accused at several points, expanded the provisions for legal aid to indigent accused persons, provided for summary trials and speedy disposal in defined classes of cases, and recognised the maintenance jurisdiction that allows a magistrate to order support for wives, children and parents who cannot maintain themselves. It preserved, deliberately, most of the procedural skeleton that generations of police officers, lawyers and magistrates already knew, because a procedural code is working infrastructure and wholesale novelty would have paralysed the courts that had to operate it from the first morning. The result is a statute that can look, to a reader comparing texts, surprisingly continuous with its predecessor, while differing from it at the two points of constitutional principle: who controls the magistrate, and whether the law offers any protection to a person before the handcuffs close. Both differences were choices, both were debated, and both have been litigated ever since.

The Forty-first Report did not appear from nowhere, and the Code’s intellectual lineage is worth a sentence because it explains the statute’s combination of conservatism and reform. The Law Commission’s Fourteenth Report of 1958, on the reform of judicial administration, had already attacked the fused magistracy and the delays of the criminal process, and the Commission’s Thirty-seventh Report of 1967 examined the 1898 Code directly before the Forty-first Report undertook the complete revision. The 1973 Code is therefore less a sudden invention than the third pass at a problem the republic had been circling since its first decade: how to keep a procedural machine that police and courts could actually operate, while rebuilding the parts of it that the Constitution had made indefensible. Parliament’s acceptance of the Commission’s two great structural recommendations, and its retention of most of the rest, set the pattern that every later amendment has followed. The architecture was treated as settled; the argument, ever since, has been about the edges, the arrest power, the victim, the timelines, where practice kept exposing the distance between the text and the station house.

How the Code of Criminal Procedure 1973 Is Built

The Code is a large working instrument rather than a short statement of principles. It contains 484 sections arranged in 37 chapters, followed by two schedules. The First Schedule classifies offences under the Indian Penal Code and other laws, stating for each whether it is cognizable or non-cognizable, bailable or non-bailable, and by which court it is triable. The Second Schedule collects the forms used in criminal process, from summonses and warrants to bonds and charges. Between the chapters and the schedules, the Code covers the constitution of criminal courts, the powers of police to investigate, the processes that compel appearance, the framing of charges, the conduct of trials in their several forms, bail, maintenance, preventive proceedings, appeals, references and revisions, and the execution of sentences. A reader who wants to know how a criminal case actually runs in India is, in substance, asking for a guided tour of this text.

How many sections does the Code of Criminal Procedure 1973 contain?

The Code of Criminal Procedure 1973 contains 484 sections arranged in 37 chapters, together with two schedules. The First Schedule classifies offences as cognizable or non-cognizable and bailable or non-bailable and names the court that can try each; the Second Schedule sets out the standard forms, such as summonses, warrants, bonds and charges, used throughout criminal process.

The opening chapters perform the Code’s conceptual groundwork, and three distinctions planted there decide cases throughout the statute. The first is between investigation, inquiry and trial, three words the Code defines separately because different powers attach to each. Investigation is the stage conducted by the police, or by a person authorised by a magistrate, to collect evidence. Inquiry is a proceeding conducted by a magistrate or court that is not a trial. Trial, a word the Code uses constantly without defining exhaustively, is the judicial determination of guilt or innocence that follows. The second distinction is between cognizable and non-cognizable offences. In a cognizable offence, a police officer may arrest without a warrant and may investigate without a magistrate’s permission; in a non-cognizable offence, the officer has neither power, and both arrest and investigation require the authority of a court. The third distinction is between bailable and non-bailable offences, which determines whether release on bail is a right the police or court must honour on demand or a discretion to be argued for. These classifications are not decorative taxonomy. The First Schedule’s entries for each offence fix, in advance and for every police station in the country, how a case begins, whether an arrest can be immediate, and what an accused person can claim by right at the station door. Almost every practical question in criminal procedure is answered by first locating the offence in that grid.

The trial chapters then divide cases by gravity. A warrant case, meaning a case relating to an offence punishable with death, imprisonment for life or imprisonment for a term exceeding two years, follows the fuller trial procedure, while a summons case, meaning any case that is not a warrant case, follows a lighter one, and sessions trials, reserved for the gravest offences, follow the most elaborate procedure of all before the Court of Session. Layered over this sits the hierarchy of criminal courts: the High Court and the Court of Session at the top of the district structure, then Judicial Magistrates of the first and second class and Chief Judicial Magistrates, with Metropolitan Magistrates in the metropolitan areas, and, on the separate executive track, District Magistrates, Sub-divisional Magistrates and Executive Magistrates exercising the Code’s preventive and administrative powers. Sentencing powers climb with the hierarchy, so the classification of the offence determines not only how the case is tried but how much punishment the trying court may lawfully impose.

Does the Code of Criminal Procedure 1973 apply to offences under special laws?

Yes, as the default procedure, because Section 4 provides that offences under the Indian Penal Code and under other laws are investigated, inquired into and tried according to the Code. Section 5 saves any procedure a particular statute prescribes for itself, so where a special law is silent the Code fills the gap, and where it differs, its procedure prevails.

The Code is therefore best understood as the general law of criminal process into which special statutes plug their exceptions. Laws against corruption, narcotics, terrorism and economic offences repeatedly borrow the Code’s machinery for investigation, remand and trial while modifying particular points, most often the bail standard and the investigation period, and the courts read the two texts together unless the special statute displaces the Code expressly or by necessary implication. This default-and-exception design is why mastery of the 1973 Code remained the foundation of criminal practice even in fields dominated by special legislation, and why arguments about procedure so often resolve into a question about whether the special statute has actually said something different or has merely been assumed to have done so.

The Separation of the Judiciary from the Executive

The separation the 1973 Code enacted was institutional, and its mechanics are worth stating precisely because the phrase is often used more loosely than the statute warrants. The Code divides magistrates into two services with two different masters. Judicial Magistrates, including Chief Judicial Magistrates and Metropolitan Magistrates, are appointed by the High Court, and the High Court exercises control over them; it is Judicial Magistrates who conduct inquiries and trials, take cognizance of offences, authorise detention during investigation, grant or refuse bail and pass sentences. Executive Magistrates, including the District Magistrate and Sub-divisional Magistrates, are appointed by the state government and remain within the executive hierarchy; to them the Code assigns the functions that are administrative or preventive in character, such as orders to disperse unlawful assemblies, proceedings against public nuisances, security proceedings for keeping the peace, and the prohibitory orders made under Section 144. The Court of Session, presided over by a Sessions Judge appointed by the High Court, tries the gravest cases. The design takes the functions in which the state appears as prosecutor and removes them from the chain of command of the state as administrator, while leaving the functions in which the state acts as keeper of public order inside that chain, where democratic accountability for order is thought to belong.

The constitutional warrant for the design is Article 50 of the Constitution, a directive principle of state policy which provides that the state shall take steps to separate the judiciary from the executive in the public services of the state. Directive principles are not enforceable by any court, and for the first two decades of the republic the separation remained substantially unachieved in the magistracy, whatever progress was made at higher levels of the judiciary. The 1973 Code is the instrument that finally carried Article 50 into the district criminal courts. Its enactment is therefore a constitutional event as much as a procedural one: the point at which a promise made in 1950 was performed in the part of the justice system that ordinary citizens actually meet, because the magistrate’s court, not the High Court, is where the criminal law touches most lives.

The honest assessment, which the brief for this article insists on and which the record supports, is that the separation was enacted more completely on paper than it was completed in fact. Executive Magistrates retained substantial and genuinely judicial-looking powers, particularly in the preventive chapters: the power under Section 144 to prohibit assemblies and acts in anticipation of danger, security proceedings that can require a person to execute a bond for good behaviour on pain of imprisonment, and proceedings over disputes concerning land and water that are likely to cause a breach of the peace. These powers can restrict liberty and property on executive satisfaction, with the affected person’s remedy lying in revision or in the constitutional courts rather than in an appeal within a judicial hierarchy. The separation also arrived unevenly across the states, at different speeds and with different degrees of completeness in staffing the new judicial magistracy, so that for years the lived reality in many districts lagged behind the Code’s scheme. Critics of the incomplete separation argue that preventive jurisdiction is precisely where executive control is most dangerous, because it operates before any offence and on prediction rather than proof. Defenders of the arrangement answer that preventive action is an executive responsibility inseparable from the duty to keep order, that the District Magistrate must answer politically when order collapses, and that the constitutional courts have kept the preventive chapters within bounds by insisting on their temporary and exceptional character. Both arguments identify something real. The Code’s settlement was a compromise: it judicialised the trial and the custody decision, which are the points at which the state accuses, and left the preventive functions executive, where the state anticipates. Whether that line was drawn at the right place remains a live argument in Indian public law, and it is an argument the Code’s structure invites every time a prohibitory order is challenged.

Arrest: Sections 41 and 41A and the Arnesh Kumar Discipline

Arrest is the point at which the criminal process takes hold of a person’s body, and the Code regulates it in unusual detail because it is also the point at which the process is most easily abused. The conceptual frame matters here: an arrest under the Code is a step inside a prosecution for an alleged offence, with charge, bail and trial ahead of it, which distinguishes it from preventive detention, a custody imposed without any offence at all; the two regimes are compared directly in the guide to preventive detention versus ordinary arrest, and the distinction controls which safeguards apply from the first hour.

Section 41 states when a police officer may arrest without a warrant. Its best-known limb covers cognizable offences, and after the amendments of 2008 and 2010 it does so in a form that repays close reading. For an offence punishable with imprisonment of up to seven years, the officer may arrest only if he has reason to believe, on the basis of a complaint, credible information or reasonable suspicion, that the person has committed the offence, and is further satisfied that arrest is necessary for one of the purposes the section lists: to prevent the person from committing a further offence, for proper investigation, to prevent the disappearance or tampering of evidence, to prevent the person from threatening or influencing witnesses, or to ensure the person’s presence in court. The officer must record the reasons for arrest in writing, and, in every case where arrest is not required, must equally record the reasons for not arresting. For offences punishable with more than seven years, arrest may rest on credible information and reason to believe. Around this core the Code builds the arrest’s surrounding discipline: Section 41B requires the arresting officer to bear visible identification and to prepare an arrest memo attested by a witness and countersigned by the arrested person; Section 41C requires police control rooms and public display of the names of arrested persons; Section 41D recognises the arrested person’s right to meet an advocate during interrogation, though not throughout it; and Section 46 and the provisions that follow regulate how the arrest is physically made and require production before a magistrate without unnecessary delay. The Supreme Court’s guidelines in D K Basu v State of West Bengal, delivered in 1997, supplied much of the arrest-memo and notification discipline before the legislature wrote parts of it into the Code, and the case remains the standard citation for the proposition that the power to arrest carries enforceable duties toward the person arrested.

The arrested person’s own rights at the moment of arrest are stated in a compact group of provisions that practitioners test first in any custody challenge. Section 50 requires the person arresting without warrant to communicate forthwith the full particulars of the offence for which the person is arrested or other grounds of the arrest, and, where the offence is bailable, to inform the person that he is entitled to be released on bail and may arrange for sureties; an arrest made without communicating the grounds is not a technical irregularity but a failure of the condition on which the arrest’s legality rests, and it engages the parallel guarantee in Article 22(1) of the Constitution. Section 50A, inserted by the amendment of 2005, obliges the person making the arrest to inform a friend, relative or other person nominated by the arrested person, as soon as practicable, that he has been arrested and where he is detained, with the particulars entered in the station book and the magistrate informed when the person is brought before him. Section 54 entitles a person arrested to be examined by a medical officer, and, as strengthened by the amendment of 2008, requires the examination of an arrested person by a registered medical practitioner soon after arrest, recording injuries and complaints, so that a contemporaneous medical record exists against which any later allegation of custodial violence can be tested. Section 55A, also a deposit of the 2008 amendment, makes it the duty of the person having custody of an accused to take reasonable care of his health and safety. Read together, these provisions answer the question the arrest power never asks itself, namely how anyone outside the station will know that an arrest has happened and in what condition the arrested person entered custody; the Code’s answer is a paper trail of communicated grounds, notified relatives and a medical record, each element of which can be demanded at the first remand hearing.

The companion provision is Section 41A, the notice of appearance, which is the Code’s statutory answer to unnecessary arrest. Where the arrest of a person is not required under Section 41, the police officer shall issue a notice directing the person against whom a complaint has been made, credible information has been received or reasonable suspicion exists to appear before the officer or at another specified place. A person who complies with the notice and continues to comply shall not be arrested for the offence mentioned in it unless the officer, for reasons recorded, forms the opinion that arrest has become necessary. Non-compliance, or unwillingness to identify oneself, can lead to arrest subject to any order of a competent court. The provision was inserted by the Code of Criminal Procedure (Amendment) Act 2008, but its commencement tells its own story about legislative hesitation: most of that Act came into force on 31 December 2009, while the arrest provisions were held back and notified only from 1 November 2010, and a further amendment in 2010 converted the officer’s power to issue the notice into a duty by substituting “shall” for “may”. The sequence matters analytically. Parliament created the notice as the default route for the large class of cases in which custody adds nothing to investigation, surrounded it with recorded reasons either way, and thereby made arrest a decision that must be justified rather than a reflex that must be excused.

What happens at a first remand hearing after arrest?

At the first production, which must occur within twenty four hours of arrest excluding journey time, the magistrate examines the case diary, hears the accused, and decides whether custody is justified. The magistrate may authorise police custody or judicial custody within Section 167 limits, release the accused on bail, or refuse custody where the material does not support detention.

It was the gap between this text and station-house practice that produced Arnesh Kumar v State of Bihar in 2014. The case arose from an arrest in a matrimonial offence, but the Supreme Court used it to address routine arrest across the class of offences punishable with up to seven years. The Court held that arrest in such cases is not automatic on the registration of a case; the police officer must apply the Section 41 checklist and be able to show, from recorded reasons, why custody was necessary, and the magistrate authorising detention must not do so casually or mechanically but must record satisfaction that the arrest itself was justified under the checklist. The judgment directed that police officers be furnished with the checklist of the sub-clauses of Section 41(1)(b)(ii), that the magistrate record reasons when authorising detention, and that a decision not to arrest be forwarded to the magistrate within the period the judgment specified, with departmental action and contempt consequences contemplated for officers who arrest without justification and for magistrates who authorise detention mechanically. Read together, Sections 41 and 41A and Arnesh Kumar state a complete philosophy: arrest is an investigative tool of last resort for offences of moderate gravity, the default instruments are notice and recorded reasons, and every departure from the default must leave a paper trail that a magistrate can audit. The philosophy’s weakness, taken up in the assessment at the end of this article, is that a checklist can be filled in as mechanically as the detention it was meant to discipline, and the judgment’s own directions depend on the same institutions whose habits they were meant to change.

The First Information Report and the Opening of Investigation

A criminal case under the Code begins, in the ordinary course, with information. Section 154 requires that every information relating to the commission of a cognizable offence, given to an officer in charge of a police station, be reduced to writing, read over to the informant, signed by the informant and entered in the book kept for the purpose at the station; a copy must be given free of cost to the informant. The resulting document, the first information report, is not evidence of the facts it records and is not expected to be an encyclopedia of the case; its legal function is to set the criminal law in motion and to fix, at the earliest moment, the version of events on the basis of which the investigation proceeds. Because so much later turns on it, the questions whether the police may refuse to register it, and what follows if they do, generated some of the most consequential litigation under the Code.

The Constitution Bench in Lalita Kumari v Government of Uttar Pradesh, which delivered its decision in November 2013, settled the first question in terms that admit little argument. Registration of a first information report is mandatory if the information discloses the commission of a cognizable offence; the officer has no discretion to conduct a roving inquiry into the truth of the allegation before registering it, because the statutory scheme assigns the testing of the allegation to the investigation that follows registration, not to a threshold screening that the Code does not authorise. The Bench permitted a preliminary inquiry only in a limited set of categories, named in the judgment, where the information may not disclose an offence on its face or where the character of the dispute requires care before the criminal process is set in motion: matrimonial and family disputes, commercial offences, cases of medical negligence, corruption cases, and cases involving abnormal delay in initiating prosecution, the list being illustrative rather than exhaustive. Even in those categories the inquiry is confined to whether a cognizable offence is disclosed, must be time bound, and does not extend to verifying the veracity of the allegation. The holding converted what had been a contested administrative practice, in which stations screened complaints for credibility, seriousness or convenience, into a rule of law: disclose a cognizable offence and the report must be recorded, after which the investigation, not the station door, is where weak allegations are filtered out.

If the officer in charge refuses, the Code does not leave the informant without a route. Section 154(3) allows the substance of the information to be sent in writing to the Superintendent of Police, who, if satisfied that a cognizable offence is disclosed, must either investigate personally or direct an investigation. Section 156(3) empowers a magistrate to order the police to investigate a cognizable case, and Section 200 allows the informant to file a complaint directly before a magistrate, who will examine the complainant on oath and may inquire or direct an investigation before deciding whether to take the matter forward. The Supreme Court in Sakiri Vasu v State of Uttar Pradesh, decided in 2008, mapped this hierarchy of remedies and held that a person aggrieved by non-registration should ordinarily work through the Code’s own ladder, to the Superintendent and then to the magistrate, rather than invoking the writ jurisdiction of the High Court as a first resort, because the Code’s remedies are efficacious and the magistrate’s power under Section 156(3) includes monitoring the investigation he has ordered to the extent the judgment describes. In practice the magistrate’s complaint route is the remedy that matters, because it transfers the initiative from the police station that has refused to a court that can compel. Two cautions from the decided cases complete the picture. The magistrate acting under Section 156(3) orders an investigation; he does not thereby take cognizance, which happens later and by a distinct act. And the powers of the police to investigate a cognizable offence under Section 156(1), within the local limits and without any magistrate’s leave, remain the ordinary engine of the system, with Section 155 performing the opposite function for non-cognizable cases, in which the police may neither investigate without a magistrate’s order nor arrest without a warrant.

Two further features of the investigative stage deserve notice because they are frequent sources of error. The first is the territorial problem and its informal solution. Information is sometimes given at a station that lacks territorial jurisdiction over the place of the offence, and the practice grew of recording the report nonetheless, numbering it as a zero first information report and transferring it to the competent station, so that the recording of the information is not defeated by a boundary dispute at the moment a victim most needs the machinery to move. The practice rested on the principle the courts articulated that the duty to record attaches to the information, not to the accident of jurisdiction, and it coexisted with the rule that an investigation by a station lacking jurisdiction is not, without more, a nullity. The second feature is the discipline of Section 157 and the provisions around it: the officer who starts an investigation must forthwith send a report to the magistrate empowered to take cognizance, and the investigation must proceed under Chapter XII of the Code, whose provisions on search, examination of witnesses and the recording of statements supply the raw material that the rest of the process will test. The investigation is the Code’s least supervised stage and its most consequential one, since the quality of everything later, charge, trial and appeal, is capped by the quality of the evidence gathered here; that asymmetry between consequence and supervision is the standing structural criticism of the design, and no amendment has removed it.

One doctrine of the investigation stage deserves separate statement because it decides trials without appearing in any section heading: the treatment of delay in lodging the first information report. The report’s function, as the courts have explained across a long line of cases, is to set the law in motion at the earliest point and to preserve the first version of events before memory is edited by consultation, reflection or interest. An unexplained delay in reaching the police station therefore invites the argument that the report is the product of deliberation rather than of the event, and trial courts weigh that argument seriously, particularly where the delay is long, the informant’s movements are unexplained, and names appear in the report that the earliest account would not have contained. Delay is not, however, a ground of automatic rejection. Where the explanation is credible, the distance, the hour, the condition of the victim, the priority of medical aid over report writing, the courts accept the report and test its contents like any other material, because the doctrine protects against embellishment and not against the realities of how news of a crime travels. The practical teaching is characteristic of the Code’s working method: the statute fixes the duty to record and the form of recording, and the courts, case by case, supply the standards by which the product of that recording is judged.

Search, Seizure and the Coercive Tools of Investigation

Investigation under the Code is not confined to asking questions. The statute arms the police and the magistracy with a graded set of coercive instruments for obtaining things, documents and access, and it surrounds each with conditions whose breach is the staple of defence argument at trial. Section 91 empowers an officer in charge of a police station, and any court, to require the production of any document or thing considered necessary or desirable for the purposes of an investigation, inquiry or trial, by summons to the person in whose possession or power it is believed to be. The power does not extend to compelling an accused person to produce incriminating material in violation of the protection against self incrimination in Article 20(3) of the Constitution, a boundary the Supreme Court drew in State of Bombay v Kathi Kalu Oghad in 1961 in distinguishing the production of documents and specimens from testimonial compulsion. Production notices under Section 91 are the everyday instrument by which bank records, registers, files and electronic records are gathered in economic offence investigations, and arguments about their width, whether the thing demanded is necessary or desirable for the proceeding, recur wherever the notice is used as a fishing instrument rather than as a targeted one.

Search powers follow the same graded logic. A court may issue a search warrant under Section 93 where a document or thing is not produced in answer to a summons, where the person in possession is unknown, or where the purposes of an inquiry or trial will be served by a general search or inspection; Section 94 addresses places suspected of containing stolen property, forged documents or objectionable material. The police possess a parallel power under Section 165 to search during investigation within the limits of their station, on reasonable grounds recorded in writing, where they believe anything necessary for the investigation may be found and cannot otherwise be obtained without undue delay; the grounds must be recorded, copies sent to the nearest magistrate, and the occupant of the place searched is entitled to a copy, because the section’s design substitutes an immediate written record for prior judicial permission. Section 100 supplies the safeguards that make a search defensible: the place is to be searched in the presence of two or more independent and respectable inhabitants of the locality, the occupant or his nominee may attend, a list of things seized is to be prepared and signed by the witnesses, and a copy delivered to the occupant. Trial courts test searches against these requirements with real consequences, since a search conducted in defiance of them damages the recovery evidence built upon it, although the decided cases stop short of holding that every irregularity automatically destroys the evidence recovered, which is assessed like other evidence with the irregularity weighed in the balance.

Section 102 completes the toolkit by empowering a police officer to seize any property which may be alleged or suspected to have been stolen, or which may be found under circumstances creating suspicion of the commission of any offence, reporting the seizure forthwith to the magistrate. The Supreme Court held in State of Maharashtra v Tapas D Neogy, decided in 1999, that a bank account is property within the meaning of the section, so that an investigating officer may direct a bank not to operate an account that forms part of the investigation’s subject matter; the freezing of accounts in economic investigations rests on that reading. The magistrate’s powers under Section 451 and Section 457 govern the custody and disposal of property produced before the court or reported as seized, and the interim release of seized vehicles and goods on bond is one of the most frequent small litigations of the criminal courts. Taken as a whole, the search and seizure provisions show the Code’s characteristic method in its purest form: the power is given in wide terms, because investigation cannot be conducted at arm’s length from physical evidence, and the power is then disciplined by writing, by witnesses, by reporting and by judicial control of the proceeds, so that the legality of the search can be reconstructed at the trial from the paper it generated.

Statements During Investigation and the Special Position of Section 164

The Code treats statements made during investigation with a deliberate and instructive suspicion. Under Section 161, a police officer investigating a case may examine persons acquainted with the facts and reduce their statements to writing, and the person examined is bound to give true answers, apart from questions whose answers would expose him to a criminal charge. Under Section 162, however, a statement so recorded may not be signed by the maker and may not be used as evidence at the trial for any purpose except the narrow one of contradicting the witness in the manner the law of evidence permits. The design is a considered judgment about institutional trust. Police recording in the station is treated as a working tool of investigation, valuable for directing the inquiry and for testing the witness who later departs from it, but too exposed to pressure, abbreviation and suggestion to be admitted as proof of what it asserts. The evidentiary rules that govern what a court may actually rely on are the province of the Indian Evidence Act 1872 and its treatment of confessions and witness statements, and the boundary between the two statutes is nowhere more visible than here: the Code gathers the material, the Evidence Act decides what the material is worth.

Section 164 creates the Code’s controlled exception to that suspicion. A Judicial Magistrate or Metropolitan Magistrate, whether or not he has jurisdiction in the case, may record a confession or a statement made in the course of investigation, and a statement so recorded can be used at trial in ways a police-recorded statement cannot. Because a confession recorded by a magistrate can decide a trial, the section surrounds the recording with safeguards that the courts have enforced strictly. Before recording a confession the magistrate must explain to the person that he is not bound to confess and that if he does, the confession may be used as evidence against him; the magistrate must satisfy himself that the confession is voluntary, must give the person time for reflection away from police influence, must not record a confession he does not believe to be voluntary, and must append a memorandum certifying the process. A confession recorded in breach of these protections invites exclusion, and the case law treats the section’s requirements as substantive rather than ceremonial, because the whole point of routing the statement through a magistrate is to break the chain of custody between the police and the words. Section 164 is also used, without any confession involved, to record the statements of important witnesses whose evidence the prosecution wants preserved in an authoritative form, and the 2013 amendments strengthened this use in the cases where it matters most: Section 164(5A), inserted by the Criminal Law (Amendment) Act 2013, requires a magistrate to record the statement of a victim of specified sexual offences as soon as the offence is brought to notice, with assistance for persons with disabilities as the provision prescribes. The trajectory across these provisions is consistent. The Code assumes that words obtained in custody are dangerous, builds one carefully guarded channel through which some of them may acquire evidentiary status, and leaves the rest to be proved in open court by witnesses who can be cross-examined.

A companion provision deserves mention here because it applies the same philosophy to the body rather than to words. Section 164A, inserted by the amendment of 2005, governs the medical examination of the victim where an offence of rape or attempt to rape is under investigation: the examination is to be conducted by a registered medical practitioner employed in a government or local authority hospital, or in his absence by any other registered medical practitioner, with the consent of the woman or of a person competent to give consent on her behalf, and the report must record the particulars the section enumerates, including the name and address of the woman and of the person who brought her, her age, the description of material collected for analysis, marks of injury and her general mental condition, together with the reasons for each conclusion. The section requires the report to be forwarded without delay to the investigating officer and onward to the magistrate, and it records expressly that nothing in it renders lawful any examination without consent. The design matches Section 164’s precisely: a process point at which the state’s handling of a vulnerable person generates evidence of lasting consequence is surrounded by consent, recording and time discipline, because the quality of the record made in the first hours determines what can be proved years later.

Remand and Default Bail: The Clock of Section 167

Section 167 is where the brief’s clock metaphor stops being a metaphor. When a person is arrested and it appears that the investigation cannot be completed within twenty four hours, the officer must transmit the entries in the case diary to the nearest Judicial Magistrate and produce the accused before him, and the magistrate may authorise detention in such custody as he thinks fit for a term not exceeding fifteen days in the whole at the first instance. Police custody, the form in which the accused remains available for interrogation, is confined to the first fifteen days after production; detention beyond that window, if authorised at all, is judicial custody, in which the accused is lodged in prison under the court’s authority and the police may question him only as the court permits. Thereafter the magistrate may extend judicial custody in stretches, but the total period of detention during investigation is capped by the proviso that drives the section’s practical importance: where the investigation relates to an offence punishable with death, imprisonment for life or imprisonment for a term of not less than ten years, the outer limit is ninety days, and in any other case it is sixty days. On the expiry of the applicable period, if the investigation is not complete and the police report has not been filed, the accused shall be released on bail if he is prepared to and does furnish bail, and the release is deemed to be under the Code’s bail chapter for purposes such as cancellation. Section 167(5) adds a further discipline at the other end of the gravity scale: in a summons case, if the investigation is not concluded within six months from the date of arrest, the magistrate shall stop the investigation unless the officer satisfies him that special reasons and the interests of justice justify continuing it.

What is a charge sheet and what does it do?

A charge sheet is the police report under Section 173 by which the investigation reaches the magistrate, naming the accused and the offences alleged. Its filing ends the investigation and hands the case to the court. If it arrives after the Section 167 period has expired, it comes too late to defeat a default bail right claimed in time.

The courts have built the doctrine of default bail on this proviso, and its shape is now settled. The right that accrues on the expiry of the sixtieth or ninetieth day, if no charge sheet has been filed, is an indefeasible right, in the word the Supreme Court used in Sanjay Dutt v State through CBI in 1994 and repeated in the line of cases that followed, including Uday Mohanlal Acharya v State of Maharashtra in 2001: it is part of the Article 21 guarantee, it cannot be defeated by the seriousness of the accusation, and a magistrate who refuses it because the offence is grave has misunderstood the provision. But the right must be availed of. If the accused does not apply for default bail after the period expires and before the charge sheet is filed, and the charge sheet is then filed, the right is extinguished and the accused is remitted to an application for bail on merits; the race between the application and the filing is therefore a real race, run in registries and courtrooms on the sixty-first and ninety-first days, and it is the purest illustration in Indian law of procedure operating as substance. The later cases have refined the edges: Rakesh Kumar Paul v State of Assam, decided in 2017, construed the ninety day category by reference to offences for which the minimum punishment prescribed is not less than ten years, so that an offence punishable with imprisonment that may extend to ten years but carries a lesser minimum falls in the sixty day class; and the Court has insisted that the right, once accrued and claimed, is not defeated by the filing of an incomplete or preliminary report designed to stop the clock, because the report that counts is the report of a completed investigation under Section 173. Two consequences follow for practice. Custody during investigation is a wasting asset for the prosecution, which must either finish its work inside the period or justify the accused’s continued detention through the ordinary bail contest after charge. And the magistrate at remand is not a rubber stamp for the investigation’s timetable but the keeper of the clock, charged with counting the days, recording the form of custody and releasing the accused when the statute commands it, whether or not release suits the investigation.

One refinement of the fifteen day rule matters in daily practice. In Central Bureau of Investigation v Anupam J Kulkarni, decided in 1992, the Supreme Court held that the fifteen days of permissible police custody are fifteen days in the whole, counted from the first production, and that within that opening window the magistrate may authorise police custody in spells, returning the accused to judicial custody between them, but that once the window closes no police custody can be authorised for that investigation, the bar being designed to end the interrogation phase of detention decisively rather than to leave it available whenever the police feel the need. The ruling interacts with the default bail clock in a way practitioners plan around: an agency that wants both spells of custody and a completed investigation must run its interrogation early and its paperwork fast, because the two clocks, the fifteen day interrogation window and the sixty or ninety day filing deadline, run together and neither waits for the other. The accused’s corresponding strategy is written into the same arithmetic, since every day of the window consumed is a day closer to the deadline at which custody itself expires by default.

The remand hearing is also where the Code’s two hands visibly meet. The prosecution’s interest is in time, and the accused’s interest is in the calendar, and the magistrate sits between a case diary that always seems to require a few more days and a liberty interest that the Constitution prices very high. The decided cases on mechanical remand, of which Arnesh Kumar is the most cited on the arrest side, insist that authorisation of detention is a judicial act requiring the application of mind to the material, not an administrative endorsement of the police request; a remand order that recites no satisfaction at all, or that extends custody after the statutory ceiling has passed, is unlawful however strong the underlying suspicion. It is worth pausing on why Parliament chose this architecture at all. The alternative, familiar from preventive detention law, is custody on executive satisfaction reviewed after the event. For the ordinary criminal process the Code rejected that model: custody belongs to courts, in short judicially supervised instalments, with a hard outer limit that converts investigative delay into release. That choice, more than any single section, is the Code’s theory of liberty in operation.

The Police Report, Cognizance and the Complaint Route

When the investigation ends, Section 173 takes over. On completion, the officer in charge must forward to the magistrate a report in the prescribed form, stating the names of the parties, the nature of the information, the names of the persons who appear acquainted with the circumstances, whether any offence appears to have been committed and if so by whom, and whether the accused has been arrested or released. Where the evidence is insufficient, the report says so and the police seek the closure of the case; where the evidence supports prosecution, the report is what practice calls the charge sheet. The filing of this report is the hinge of the whole Code. Before it, the case belongs to the police under judicial supervision of custody; after it, the case belongs to the court, and the police become witnesses and assistants in a process they no longer control. Section 173(8) preserves the possibility of further investigation after the report is filed, with the court’s leave in the situations the case law describes, so the hinge is not a sealed door, but the transfer of initiative it effects is real and is the structural answer to the fear that an investigation, once closed by its own conductor, can never be revisited.

Between investigation and trial stands cognizance, the court’s formal act of applying its mind to the suspected commission of an offence with a view to initiating proceedings. Section 190 empowers a magistrate to take cognizance in three ways: upon receiving a complaint of facts constituting the offence, upon a police report of such facts, and upon information received from any person other than a police officer or upon his own knowledge. Cognizance is taken of the offence, not of the offender, and it is not a finding that anyone is guilty; it is the decision that the material justifies setting the trial machinery in motion, after which the accused is summoned and the case enters the trial chapters. On a police report seeking closure, the magistrate is not bound by the police conclusion. The Supreme Court in Bhagwant Singh v Commissioner of Police, decided in 1985, held that the informant is entitled to notice and an opportunity of being heard before the magistrate accepts a final report and closes the case, because the informant set the law in motion and has a legitimate interest in its outcome. If the informant appears and contests the closure, the statement of objections is known in practice as a protest petition, and the magistrate faced with it has options the case law maps: accept the closure notwithstanding the protest, treat the protest as a complaint and proceed under the complaint procedure, or disagree with the police and take cognizance on the material gathered. The informant who began the case at the police station thus reappears at its hinge, armed by the Code with a right to be heard before the case he started is buried.

Cognizance is also the point at which the Code places its statutory screens, of which Section 197, treated later in this article, is the best known but not the only one. Section 195 bars courts from taking cognizance of offences against public justice, of offences against the lawful authority of public servants and of connected document offences, except on the complaint in writing of the public servant or the court concerned, a design that prevents the criminal process from being used to retaliate against officials and courts for acts done in their functions while preserving a channel for genuine complaint through the authority itself. Section 196 requires the previous sanction of the Central or state government, as the case may be, for cognizance of offences against the state, of criminal conspiracy to commit them, and of the class of offences promoting enmity between groups and outraging religious feelings that the section specifies, offences in which Parliament judged that prosecution itself carries public consequences a government should answer for before the process begins. Together with Section 197, these provisions mark the class of cases in which the Code distrusts private initiative at the threshold, not because the offences are minor but because the decision to prosecute them is partly a decision about public order and institutional functioning, which the Code assigns to governments that can be questioned for it politically in a way a private complainant cannot.

The complaint route, which Sections 200 to 203 govern, is the Code’s standing alternative to police initiative, and it matters far beyond the refusal-to-register situation in which it first appeared in this article. Any person may set the criminal law in motion by complaining to a magistrate, without the police as intermediary. The magistrate receiving the complaint examines the complainant and any witnesses present on oath, reduces the substance to writing, and then decides: he may inquire into the case himself or direct an investigation by the police or another person under Section 202, may postpone the issue of process while doing so, must dismiss the complaint under Section 203 if he finds no sufficient ground for proceeding, and issues process under Section 204 if he finds such ground. The procedure disciplines private prosecution without disabling it. The oath and the preliminary inquiry filter the frivolous and the vindictive, while the route itself guarantees that no police station and no local power structure can confer immunity on an offender by simply declining to act, because the court door stands open behind the station door. A great deal of Indian criminal litigation, in fields from defamation and cheque dishonour to offences within families and businesses, runs on this track rather than on police reports, and the Code’s design treats the two tracks as equivalents feeding the same trial process.

Processes to Compel Appearance: Summonses, Warrants and Proclamation

Once cognizance is taken, the court’s first practical problem is getting the accused before it, and the Code supplies an escalating ladder of process. The ordinary instrument is the summons, whose form and service occupy Sections 61 to 69: a summons must be in writing, in duplicate, signed by the presiding officer and bearing the seal of the court, and it is served personally where practicable, with service on an adult member of the family, by affixation, or on a corporation or government servant through the channels the later sections prescribe. Where the court sees reason, it may instead issue a warrant of arrest under Sections 70 to 81, which must likewise be in writing and signed, remains in force until executed or cancelled, and may be endorsed for execution anywhere in India, with the person arrested to be produced before a magistrate without unnecessary delay and in any case, subject to the bail provisions, within twenty four hours. Section 87 allows the court to issue a warrant in lieu of, or in addition to, a summons where the court has reason to believe the accused has absconded or will not obey the summons, or where he has failed to appear and the summons is proved to have been duly served. Section 88 empowers the court to take a bond, with or without sureties, for the appearance of a person present in court or brought before it, converting custody into an obligation recorded in advance.

The ladder’s top rungs answer the accused who will not come at all. Section 82 provides for a proclamation where the court has reason to believe that a person against whom a warrant has been issued has absconded or is concealing himself so that the warrant cannot be executed: the proclamation, requiring the person to appear at a specified place and time not less than thirty days from its publication, is publicly read in the place where the person ordinarily resides, affixed to a conspicuous part of his house or homestead and to the court house, and may be published in a newspaper. Section 83 permits the court, after issuing the proclamation, to order the attachment of the absconder’s property, which may be sold if the person does not appear within the period the section prescribes, and Section 84 and the provisions around it govern claims and appeals in respect of attached property and its restoration when the person appears or is apprehended. The scheme is coercive in a carefully economic way: it does not punish the absconder for an offence he has not been tried for, but it makes flight progressively more expensive, and it brands the status formally, since a person against whom a proclamation has issued is described in the record and in later proceedings as a proclaimed offender, with consequences in bail and in any fresh arrest. The jurisprudence on these sections is mostly about discipline at the margins, proclamations issued without the court recording its satisfaction, attachments made without observing the sequence the sections require, and the courts have repeatedly held that the coercive ladder must be climbed in order, each rung justified by the failure of the one below it.

The Trial: Charge, Evidence, the Accused’s Statement and Judgment

The trial chapters are the Code’s longest sustained piece of design, and their pivot is the charge. Sections 211 to 224 prescribe what a charge must contain: the offence with which the accused is charged, the law and the section of the law against which the offence is said to have been committed, and particulars of the time, place and person against whom the offence was committed that are reasonably sufficient to give the accused notice of the matter with which he is charged. The charge is read over and explained to the accused, who is asked whether he pleads guilty or has a defence to make. Errors and omissions in the charge do not vitiate it unless the accused was in fact misled and a failure of justice has been occasioned, and the court may alter or add to a charge at any time before judgment, reading the altered charge afresh and, where the alteration prejudices either side, adjourning or directing a new trial to the extent fairness requires. In sessions trials and in warrant cases instituted on a police report, the accused may be discharged before the charge stage if the judge or magistrate considers the charge groundless, a filter that requires the court to examine the investigation’s material and record reasons. The design choice embedded in these provisions is frequently missed: the Code does not allow a trial to drift into existence. A judicial officer must state, in a formal document the accused can hold, exactly what is alleged and under what law, and must certify at the threshold that the material warrants putting a citizen through a trial at all.

The prosecution then leads its evidence, and the witnesses are examined, cross-examined and re-examined in the manner the law of evidence governs. When the prosecution’s case closes, the court examines the accused under Section 313, a provision of distinctive character in the Code. The court puts to the accused, without oath, the circumstances appearing in the evidence against him, so that he may personally explain them; his answers are not evidence in the strict sense and he is not liable to punishment for refusing to answer or for answering falsely, but the answers may be taken into consideration in the trial and may lend assurance to other evidence, while a circumstance not put to him cannot fairly be used against him, which is the rule the courts enforce when a conviction rests on an incriminating circumstance the accused was never asked about. After the examination, the accused may enter on his defence, lead evidence, and apply for process to compel witnesses, and in a sessions trial the court must acquit at the close of the prosecution evidence if it finds there is no evidence that the accused committed the offence, without waiting for a defence that cannot cure an empty record. Two adjacent provisions complete the trial court’s instruments. Section 320 permits the compounding of the offences it lists, by the persons it names, with the permission of the court where the section requires it, ending the prosecution by composition because the offences concerned are those the law treats as substantially private wrongs. Section 321 permits the Public Prosecutor, with the consent of the court, to withdraw from the prosecution before judgment, a power whose exercise the courts supervise by requiring that the prosecutor apply an independent mind and that the consent rest on relevant considerations of public interest rather than on instruction alone.

Judgment itself is governed by Sections 353 to 363 in the trial chapters’ closing scheme and by the parallel provisions for each trial form. The judgment must be pronounced in open court by the presiding officer, must contain the point or points for determination, the decision and the reasons, and must specify the offence and the section under which the accused is convicted or the ground of acquittal. On conviction the court hears the accused on the question of sentence, a distinct stage the Supreme Court has treated as substantive rather than formal, since it is the point at which mitigation, age, antecedents, provocation and the possibility of reformation are placed before the court. Section 357 empowers the court, when imposing sentence, to order the accused to pay compensation to the victim out of any fine imposed or independently where the statute permits, and Section 360 permits the release of certain offenders on probation of good conduct or after due admonition instead of sentence. The trial’s paperwork ends, as the investigation’s began, in documents whose adequacy is appealable: a judgment that does not state reasons, or a sentence passed without hearing the accused on it, is vulnerable in the forum above, which is the Code’s final application of its ruling idea that every exercise of power in the process must be able to show its workings.

The trial chapters also contain the Code’s one sustained attempt to discipline the trial’s own duration, and its partial failure is instructive. Section 309 provides that in every inquiry or trial the proceedings shall be held as expeditiously as possible and that, when the examination of witnesses has begun, the trial shall be continued from day to day until all the witnesses in attendance have been examined, unless the court finds an adjournment necessary for reasons it records. Later amendments sharpened the discipline: the amendments of 2013 attached the day to day requirement more firmly to trials, and those of 2018 prescribed that the trial of specified sexual offences be completed, as far as possible, within two months of the filing of the charge sheet, matching the compressed investigation period the same amendment wrote into Section 173. Yet the trial stage remains the one part of the process where no default fires when the deadline passes. An investigation that overrunning produces default bail; a trial that overruns produces only an adjournment application, a complaint to the High Court’s administrative side, or an argument from the speedy trial guarantee of Article 21, which the courts have consistently declined to translate into a fixed outer limit after which prosecutions must fail. The asymmetry is the trial stage’s standing criticism, and it explains why the Code’s clock, so exact before cognizance, is felt by accused persons and victims alike to stop at the moment the trial begins.

The defence side of the trial has its own small cluster of provisions, and a pillar account should name them because they are the points at which the Code’s promise of a fair contest is cashed. Section 303 provides that any person accused of an offence before a criminal court, or against whom proceedings are instituted under the Code, may of right be defended by a pleader of his choice. Section 304 addresses the accused who has no sufficient means to engage one: in a trial before the Court of Session, where the accused is not represented and it appears to the court that he lacks the means, the court shall assign a pleader for his defence at the expense of the state, under rules the High Court frames with the state government’s approval, and the scheme may be extended to other classes of trials by state notification. Sections 306 and 307 govern the tender of pardon to an accomplice, by which a Chief Judicial Magistrate or, after committal in the situations the sections describe, the trial court may tender a pardon to a person supposed to have been directly or indirectly concerned in an offence, on condition of his making a full and true disclosure of the whole circumstances within his knowledge; the approver so created is examined as a witness, is detained until the termination of the trial unless already on bail, and forfeits the pardon if he conceals anything essential or gives false evidence. Each provision answers a structural weakness of the adversarial design the Code adopted: unequal representation, unrepresented poverty and the insider evidence without which conspiracies cannot be proved, and each was written into the Code rather than left to convention because the stakes of their absence are convictions and acquittals.

Limitation: The Code’s Deadline for Beginning a Prosecution

The clocks described so far run against the investigation and against custody; Chapter XXXVI of the Code, Sections 467 to 473, sets a clock against the prosecution itself. Section 468 bars a court from taking cognizance of an offence after the expiry of a limitation period graded by the gravity of the offence: six months if the offence is punishable with fine only, one year if it is punishable with imprisonment for a term not exceeding one year, and three years if it is punishable with imprisonment for a term exceeding one year but not exceeding three years. For offences punishable with more than three years, the chapter prescribes no bar, because Parliament judged that serious offences should not be defeated by the calendar. The period begins, under Section 469, on the date of the offence or, where its commission was unknown, on the day it first came to the knowledge of the person aggrieved or of a police officer, and the chapter excludes periods consumed by other prosecutions pursued in good faith, by absence from India and by the time taken to obtain a required sanction, among the exclusions Sections 470 and 471 enumerate. Section 473 gives the court a discretion to take cognizance after the period where the delay has been properly explained or where it is necessary in the interests of justice, a discretion the courts exercise by recorded reasons and not by routine. The chapter’s presence in the Code is the completion of its temporal scheme. A statute that disciplines the police with a charge sheet deadline and the magistrate with a remand ceiling could not consistently leave the citizen exposed forever to stale prosecutions for minor offences, and the limitation bar, modest as its periods are, is the point at which the Code’s clock protects the accused not from custody but from the accusation itself.

Bail in Three Registers: Sections 437, 438 and 439

Bail is where the Code’s theory of pre-trial liberty becomes concrete, and the statute states it in three registers that must not be confused. The first register is Section 436, bail by right in bailable offences: where a person accused of a bailable offence is arrested or appears or is brought before a court, he shall be released on bail on furnishing the required bond, because the First Schedule has already decided that offences of that class do not ordinarily justify custody. The second register is Section 437, bail in non-bailable offences as a matter of judicial discretion, exercised mainly by magistrates and structured by the section’s own signals: the gravity of the accusation, the severity of the likely punishment, the danger of the accused absconding or tampering with evidence or witnesses, and the special solicitude the section directs for persons who are sick, infirm, women or under sixteen. Section 437 also carries the section’s internal bars and conditions, including the restriction on bail where there appear reasonable grounds for believing the accused guilty of an offence punishable with death or life imprisonment, subject to the provisos the section itself contains. The third register is Section 439, the special powers of the High Court and the Court of Session in respect of bail, which are wider than the magistrate’s and include the power to impose conditions, to direct release in any case, and to cancel bail that has been granted. The Supreme Court’s general jurisprudence, restated across decades, treats bail as the rule and custody before conviction as the exception that must justify itself, a presumption rooted in the presumption of innocence and in the recognition that an undertrial who is ultimately acquitted has served a sentence no court imposed.

Who can grant anticipatory bail, and where is it sought?

Anticipatory bail under Section 438 can be granted only by the High Court or the Court of Session, on the application of a person who believes he may be arrested on an accusation of a non-bailable offence. It is ordinarily sought first in the Court of Session. A magistrate has no power under the section.

Anticipatory bail, Section 438, is the 1973 Code’s invention, and its genealogy explains its shape. The Forty-first Report of the Law Commission recommended the jurisdiction because the Commission had noticed how accusations were used as instruments: influential or hostile persons procured the implication of rivals in criminal cases, and the arrest that followed inflicted disgrace and detention that the eventual collapse of the case could not undo. The remedy had to operate before arrest precisely because the injury it targeted was the arrest itself. Section 438 therefore empowers the High Court and the Court of Session, when a person has reason to believe he may be arrested on an accusation of a non-bailable offence, to direct that in the event of arrest he shall be released on bail, on such conditions as the court thinks fit, which conventionally include making oneself available for interrogation, not influencing witnesses, not leaving India without permission, and cooperating with the investigation. The Constitution Bench in Gurbaksh Singh Sibbia v State of Punjab, decided in 1980, read the section generously and refused to straitjacket it: the discretion is to be exercised on the facts of each case, a blanket order covering offences not yet contemplated is impermissible, the filing of a first information report is not a condition of invoking the jurisdiction, and the court may impose conditions without converting the section into a copy of the post-arrest provision. The question that outlived Sibbia was duration. Successive benches divided on whether anticipatory bail should ordinarily be limited in time or should endure, and a five judge Constitution Bench in Sushila Aggarwal v State (NCT of Delhi), decided in January 2020, settled the matter: the protection granted under Section 438 should not invariably be limited to a fixed period, it does not normally come to an end when the accused is summoned or charges are framed, and it can continue till the end of the trial, though a court may limit it where the features of a case justify doing so.

Bail granted is not bail beyond recall, and the Code’s cancellation provisions complete the register. A magistrate who has released a person on bail under Section 437 may arrest him and commit him to custody under Section 437(5) if the court considers it necessary, and the High Court or Court of Session may under Section 439(2) direct that a person released on bail be arrested and committed to custody; anticipatory bail carries its own cancellation route through the court that granted it. The decided cases hold the power to a high standard, because cancellation withdraws a liberty already granted: supervening circumstances, interference with witnesses, breach of conditions or conduct obstructing the trial are the recognised grounds, while the mere filing of a charge sheet or the seriousness of the charge, without more, is not, and the Supreme Court has repeatedly distinguished the considerations for refusing bail from the stricter considerations for cancelling it. The modern restatement of the whole field is Satender Kumar Antil v Central Bureau of Investigation, decided in 2022, in which the Court consolidated its earlier directions into a compliance framework for arrest and bail, grouping offences into categories, insisting that the Arnesh Kumar discipline be actually applied by magistrates at first production, deprecating the practice of insisting on custody where the accused has cooperated and the investigation is complete, and directing that bail applications be decided promptly rather than kept pending while custody hardens. Antil is best read as the Code’s bail philosophy restated as an audit: the text had not changed, and the Court’s complaint was that the institutions were not performing it.

State variation qualifies the national picture in one important respect. The Code extends to the whole of India, but criminal procedure sits in the Concurrent List, and states have amended Section 438 locally, most famously in Uttar Pradesh, where a state amendment of 1976 omitted Section 438 altogether and anticipatory bail was simply unavailable for over four decades, until a state amendment of 2019 restored the provision in a modified form. The episode is a useful discipline against any sentence of the form “the Code provides” uttered without a territorial qualifier: the procedural rights of an accused in India have always depended partly on which state’s version of the Code governs the case, and the pillar’s account of any single provision is complete only when its state amendments are checked. What unites the registers is the principle that gives the bail chapter its constitutional weight. Each provision answers the same question at a different procedural moment, namely how much custody the state may take from a person it has not yet convicted, and the Code’s answers are graduated in a deliberate direction: by right where the offence is minor, by structured discretion where it is grave, and in advance, before the arrest happens, where the accusation itself may be the instrument of harm.

One provision added to this architecture by the amendment of 2005 deserves separate notice because it addresses the system’s most persistent scandal from within the bail chapter’s logic. Section 436A provides that where a person has, during the period of investigation, inquiry or trial, undergone detention for up to one half of the maximum period of imprisonment specified for the offence charged, he shall be released by the court on his personal bond, with or without sureties, the provision excluding offences for which death is one of the punishments specified and leaving the court a reasoned discretion to order continued detention beyond the halfway point after hearing the public prosecutor. The section caps the worst case the remand system can produce, the undertrial who spends longer awaiting trial than the sentence the offence could carry, and it converts what had been an argument from Article 21 into a statutory entitlement a magistrate can apply without waiting for a constitutional petition. Its existence on the page, and the frequency with which undertrials nonetheless remain in custody beyond its limits through want of an application, is among the sharpest illustrations in the Code of the distance between a right enacted and a right delivered.

The Inherent Powers of the High Court and Section 482

Section 482 is one of the shortest operative provisions in the Code and one of the most invoked. It provides that nothing in the Code shall be deemed to limit or affect the inherent powers of the High Court to make such orders as may be necessary to give effect to any order under the Code, or to prevent abuse of the process of any court, or otherwise to secure the ends of justice. The section does not confer the inherent powers; it saves them, recognising that a procedural code cannot anticipate every situation in which a court’s own process may be turned into an instrument of injustice, and that the High Court must retain the capacity to intervene outside the code’s express appellate and revisional channels. Petitions under Section 482 are the standard vehicle for seeking the quashing of a first information report or of criminal proceedings, for directions to protect the fairness of an investigation or trial, and for orders in the residue of situations the Code’s specific provisions do not reach.

The governing map of the quashing power is State of Haryana v Bhajan Lal, decided in 1992, in which the Supreme Court, after surveying the authorities, set out by way of illustration the categories of case in which the inherent power, or the writ power, may be exercised to quash. The categories, stated here in compressed prose, cover allegations in the first information report or complaint that, taken at their face value and accepted in their entirety, do not disclose the commission of any offence; allegations that disclose a cognizable offence but rest on material so improbable or no evidence at all; proceedings barred by an express legal provision or by the bar of a special statute; allegations that constitute an offence but are attended by a legal bar to the institution or continuance of the proceedings; and cases where the proceeding is manifestly attended with mala fides or is maliciously instituted with an ulterior motive for wreaking vengeance. The Court was careful to fence the jurisdiction as firmly as it described it. The power is to be exercised sparingly, in the rarest of rare cases in the judgment’s phrase; the High Court does not conduct a mini trial, does not weigh the evidence as the trial court will, does not quash merely because the accused has a plausible defence, and does not stifle a legitimate prosecution at the threshold. Quashing is the exception that protects the process from abuse, not an alternative forum for deciding the merits early. The balance Bhajan Lal strikes explains both the volume of Section 482 litigation and its outcomes: the door is open in every case, the threshold is deliberately high, and the petitions that succeed are those that can show, on the prosecution’s own material taken at its highest, that the proceedings should never have begun or cannot lawfully continue.

A distinct and heavily used branch of the Section 482 jurisdiction concerns settlement. Many prosecutions under the Code arise from disputes that are, in substance, private: matrimonial conflict, neighbourhood quarrels, commercial disagreements that crossed into criminal allegations. Section 320 permits the compounding of the offences it lists, but many offences charged in such disputes fall outside its table. In Gian Singh v State of Punjab, decided in 2012, the Supreme Court held that the High Court’s inherent power is not confined by Section 320: in the exercise of the Section 482 jurisdiction, criminal proceedings may be quashed on the ground of a genuine settlement where the offence is predominantly private in character and the possibility of conviction is remote once the parties have settled, because continuing the prosecution would be an exercise in futility that the ends of justice do not require. The Court drew the line at the other end with equal clarity: heinous and serious offences, offences under special statutes and offences that are public in their nature cannot be quashed merely because the victim and the accused have settled, since the wrong in such cases is not the victim’s alone to forgive. The Gian Singh settlement doctrine is Section 482 operating at its most characteristic: a reserve power, exercised on the face of the record, to stop a process whose continuation would abuse the court’s own machinery, and fenced by an insistence that the power not become a private market in public wrongs. For the pillar’s purposes, the section completes the Code’s internal safety system: the magistracy supervises custody and charge, the Sessions and High Courts hear appeals and revisions under the Code’s express provisions, and Section 482 stands behind them all as the reserve power by which the High Court polices the integrity of the process itself.

Appeals, Revision and the Supervision of the Criminal Courts

The Code’s supervisory architecture has two storeys, and the distinction between them governs strategy in every concluded case. The first storey is appeal, governed by Chapter XXIX. Section 372 states the foundation rule: no appeal lies from any judgment or order of a criminal court except as provided by the Code or by any other law, so the right to appeal is a statutory creation and must be found in a provision. The provisions that follow distribute it. Section 374 gives a person convicted on a trial held by a High Court an appeal to the Supreme Court, convicted persons in sessions trials an appeal to the High Court, and persons convicted by magistrates appeals to the Court of Session, with special routes where sentences are combined or passed by particular courts, and the chapter separately regulates appeals against acquittal, for which Section 378 requires the direction of the state government or, in a complaint case, the special leave of the High Court on the complainant’s application, a deliberate asymmetry designed to protect the finality of acquittals against routine state relitigation. The amendment of 2008 added the victim to this architecture: the proviso to Section 372 gives a victim the right to prefer an appeal against an order acquitting the accused, convicting for a lesser offence, or imposing inadequate compensation, so that the person most affected by the outcome is no longer dependent on the state’s decision to challenge it. Section 386 states the appellate court’s powers, to dismiss the appeal, to allow it and reverse the finding and sentence, to order a retrial, to alter the finding while maintaining or reducing the sentence within its competence, powers that make an appeal a rehearing on the record rather than a fresh investigation of the facts.

The second storey is revision, governed by Chapter XXX. Sections 397 to 401 empower the High Court and the Sessions Judge to call for and examine the record of any proceeding before an inferior criminal court to satisfy themselves as to the correctness, legality or propriety of any finding, sentence or order and as to the regularity of the proceedings. Revision is supervisory rather than appellate: it does not rehearse the merits at large, it does not lie against interlocutory orders, which Section 397(2) expressly excludes in order to prevent the trial from being stalled by challenges to every intermediate direction, and the revisional court may not convert a finding of acquittal into one of conviction. The powers of transfer in Sections 406 and 407, by which the Supreme Court and the High Court may move a case between courts where a fair and impartial inquiry or trial cannot be had, and the reference jurisdiction in Section 395, by which a subordinate court states a case involving a question of law for the High Court’s decision, complete the supervision scheme. The Code’s philosophy across both storeys is consistent with everything else in it: outcomes are trusted to the trial court that saw the witnesses, correction is structured and provision bound, and the higher courts intervene through stated channels for stated defects, with Section 482 held in reserve for the case that fits no channel and yet cannot be allowed to stand.

The Code Beyond the Trial Track: Maintenance, Prohibitory Orders and Sanction

A pillar account that stopped at investigation and trial would misdescribe the Code, because some of its most used provisions are not directed at punishing offences at all. Section 125, which sits in the chapter on maintenance of wives, children and parents, empowers a Magistrate of the first class, on proof of neglect or refusal, to order a person with sufficient means to pay a monthly allowance for the maintenance of a wife unable to maintain herself, of legitimate or illegitimate minor children and of adult children prevented by abnormality or injury from maintaining themselves, and of a father or mother unable to maintain themselves. The proceeding is summary and the jurisdiction is preventive rather than penal: its object, as the Supreme Court has repeatedly explained, is to prevent destitution and vagrancy by compelling support through a fast civil style remedy housed in a criminal code, enforceable by imprisonment for default but not dependent on proving any offence. The section applies across religions as a social justice measure of general application, and its litigation history, including the line of cases on divorced wives and the statutory responses those cases provoked, is one of the fullest examples in the Code of a provision whose meaning was made as much in the courts as in the legislature. For this article the point is structural: the 1973 Code is not only a trial code but the general procedural statute of the criminal courts, and Parliament repeatedly used it as the home for summary protective jurisdictions that needed a magistrate, a short procedure and an enforceable order.

Section 144 represents the preventive jurisdiction at its most exposed. It empowers a District Magistrate, a Sub-divisional Magistrate or another Executive Magistrate specially empowered, in urgent cases of nuisance or apprehended danger, to issue a written order, stating the material facts, directing persons to abstain from specified acts or to take specified order with property in their possession or management. The order is directed at situations such as unlawful assemblies, riots and serious disturbances of public tranquillity, cannot remain in force for more than two months unless the state government extends it by notification, and cannot be extended beyond six months in all; the person aggrieved may apply for rescission or alteration, and the magistrate must give an early hearing. The provision’s constitutional validity was upheld by the Supreme Court in Babulal Parate v State of Maharashtra in 1961, under the predecessor provision of the old Code, on the footing that it is a temporary, emergency power hedged by duration limits and by the duty to state material facts. Its working life has been more contentious than its validity. Orders under Section 144, which practice has converted into a general instrument for prohibiting assemblies, processions and, in the internet era, even access to communications in disturbed areas, are criticised by civil liberties advocates, among them the Internet Freedom Foundation in its challenges to prohibitory and shutdown orders, for being issued in standard form, renewed in substance beyond the section’s emergency rationale, and deployed against lawful protest. District administrations answer that the power is preventive precisely because waiting for an offence forfeits the chance to prevent violence, and that the section’s short durations and statement-of-facts requirement, honestly observed, keep it exceptional. The courts have split the difference in a recognisable way: the power stands, but orders must disclose the emergency they answer, must be proportionate to it, and must not be perpetuated by mechanical repetition, a position the Supreme Court restated in 2020 in Anuradha Bhasin v Union of India in the context of prohibitory orders and communications restrictions in Jammu and Kashmir. The tension is the separation debate in miniature, an executive power over liberty, sustained because prevention is an executive function, and policed because its exercise decides how much public life a district may conduct.

Section 144 sits within a wider preventive scheme that the pillar should not pass over, because in many districts it is the Code’s most frequently used part. Chapter VIII, Sections 106 to 124, authorises the courts and Executive Magistrates to require security for keeping the peace and for good behaviour: a person convicted of offences involving a breach of the peace may be required to execute a bond at sentencing under Section 106, and, more consequentially, an Executive Magistrate who receives information that a person is likely to commit a breach of the peace, or is a habitual offender within the descriptions Section 110 gives, may initiate proceedings by an order under Section 111 stating the substance of the information, followed by an inquiry under Section 116 conducted as nearly as may be in the manner of a summons trial, at the end of which the person may be required under Sections 117 and 118 to execute a bond, with imprisonment in default under Section 122 that can run for the period of the bond, up to three years in the habitual offender cases. Chapter X adds the powers to command unlawful assemblies to disperse, with the aid of the armed forces where the statute’s conditions are met, the proceedings under Section 133 for the removal of public nuisances by conditional order, covering obstructions, dangerous buildings, trades injurious to health and the like, and the urgent proceedings under Section 145 where a dispute concerning land or water is likely to cause a breach of the peace, in which the Executive Magistrate attaches the subject of dispute or appoints a receiver and holds possession as the civil court’s eventual determination directs. These chapters are where the Code’s criminal process shades into district administration, and they are conducted overwhelmingly by Executive Magistrates, which is why the separation debate, recorded earlier in this article, is not an antiquarian quarrel about 1974 but a description of where preventive power over liberty still sits. The preventive chapters also display the Code’s recurring safeguard pattern in its executive form: written orders stating material facts, an inquiry before the bond is finally required, fixed maximum durations, and revision to the sessions court, procedural hedges around a jurisdiction that operates, by design, on prediction rather than proof.

Section 197 performs the opposite protective function, protecting the machinery of the state against the process rather than protecting citizens against the machinery. It provides that where a judge, magistrate or public servant not removable from office save by or with the sanction of the government is accused of an offence alleged to have been committed while acting or purporting to act in the discharge of his official duty, no court shall take cognizance of the offence except with the previous sanction of the competent government, Central or state as the case may be. The rationale is that public servants must be able to take hard decisions without the paralysis of facing prosecution for each of them, and that a screening authority should test whether the accusation is a genuine complaint of crime or a reprisal for official action. The criticism, pressed by anti-corruption advocates and reflected in a long line of litigation, is that sanction becomes a bottleneck and sometimes a shield, since the government asked to sanction the prosecution of its own officers may delay or refuse for reasons that have nothing to do with the merits. The courts have answered with qualifications rather than abolition: the protection extends only to acts bearing a reasonable nexus with official duty, so that an act wholly unconnected with the office, or one in which the office is used as a mere cloak, is outside the section; the sanctioning authority must apply its mind to the material; and undue delay in deciding sanction has itself attracted judicial intervention, including directions that decisions be taken within stated periods. Section 197 is therefore best read as the Code’s most explicit institutional compromise, an admission that prosecution is itself a power that can be abused, placed at the single point, cognizance, where a screen can operate without ending the case.

The Amendment Arc: Plea Bargaining, Victims and the Rewriting of Arrest

The Code that entered force in 1974 was amended many times, but three amendments changed its character, and each is a case study in a different mode of reform. The amendment of 2005, enacted as the Code of Criminal Procedure (Amendment) Act 2005, introduced plea bargaining into Indian criminal procedure by inserting Chapter XXI-A, covering Sections 265A to 265L. An accused charged with an offence not punishable with death, life imprisonment or imprisonment exceeding seven years, and not within the categories the scheme excludes, including offences against women and children and notified socio-economic offences, may apply to the court for a plea bargain; the court examines the application, a mutually satisfactory disposition is worked out involving the prosecutor, the victim and the accused, compensation to the victim is part of the disposition, and the court sentences accordingly, with the judgment substantially final. The reform answered the docket crisis, the standing Indian condition of trials outlasting the sentences they could produce, but it was deliberately fenced: the offences eligible are the lesser ones, the victim is a participant rather than a spectator, and the disposition requires judicial approval. Its uptake in practice has been modest compared with the scale its designers hoped for, a fact its defenders attribute to institutional habit and its critics to restrictions that keep the scheme from the cases that clog the courts most.

The amendment of 2008, enacted as Act 5 of 2009, reoriented the Code toward the victim and toward the discipline of arrest, and its staggered commencement has already appeared in this article’s account of Section 41A. The bulk of the Act came into force on 31 December 2009; the arrest provisions followed on 1 November 2010, adjusted by the Amendment Act of 2010. Its substantive deposits are considerable. It inserted a definition of “victim” in Section 2(wa), a small textual event with large consequences, because once the victim is a defined person in the Code the statute can attach rights to the status. It inserted Section 357A, obliging every state government, in coordination with the Central Government, to prepare a scheme for compensating victims or their dependents who have suffered loss or injury from crime and require rehabilitation, administered through the Legal Services Authorities, so that compensation no longer depended entirely on the conviction and means of the offender. It strengthened the arrested person’s position through the arrest discipline already described, recognised the victim’s right to engage an advocate to assist the prosecution under the proviso to Section 24(8), and gave the victim a right of appeal against acquittal, conviction for a lesser offence or inadequate compensation through the proviso to Section 372. Read together, these changes mark the moment the Code’s centre of gravity visibly shifted: the statute of 1973 had been written almost entirely around the state and the accused, and the amendment of 2008 began, incompletely but irreversibly, to write the victim in as a third presence with standing of her own.

The third amendment, the Criminal Law (Amendment) Act 2013, was enacted in the aftermath of the gang rape and murder in Delhi in December 2012 and the report of the committee headed by Justice J S Verma, and it changed the Code at the points where the process had failed victims most visibly. Its procedural changes, which are analysed in full in the account of the Criminal Law (Amendment) Act 2013 and its procedural changes, included requiring the information of specified offences against women to be recorded by a woman police officer, requiring the statement of such a victim to be recorded at a place of her choice and by a woman officer where the statute prescribes, inserting Section 164(5A) for the prompt recording of victims’ statements by a magistrate, requiring investigation of specified offences to be completed within the periods the amendments prescribed, inserting Section 357C so that hospitals must provide first aid and medical treatment free of cost to victims of specified offences, and adjusting Section 197’s field in the same class of cases. The 2018 amendments to the criminal law tightened the same territory further, prescribing compressed timelines for the investigation and trial of sexual offences against children and women. The pattern across 2005, 2008, 2013 and 2018 is worth naming, because it is the Code’s late biography in one sentence: Parliament almost never reopened the Code’s architecture, but repeatedly amended its edges in response to a demonstrated failure, a docket crisis, a victim-blindness, a public atrocity, so that the statute of 2024 was the statute of 1974 in structure and a materially different instrument in its treatment of arrest, victims and sexual offences.

The Criminal Case Timeline: The Code as a Working Instrument

The provisions traced in this article operate as one sequence in a real case, and the sequence is the artifact practitioners and students actually keep. Advocates maintain the timeline as a working document, often rendered as a spreadsheet or a typed table in the case file, and a reader who keeps similar files can open such office documents in the browser with the ReportMedic office file viewer for spreadsheet, document and presentation files, which is the practical way to check a stage, a provision and a deadline side by side without specialist software. What follows is the timeline in the form the Code itself dictates, with the actor, the limit and the price of missing it stated at each stage.

Stage	Provision	Who acts	Time limit	If the limit or duty is missed	Remedy available to the accused
Information of a cognizable offence	Section 154	Officer in charge of the police station	Recorded forthwith on receipt	Registration is mandatory where a cognizable offence is disclosed, per Lalita Kumari	Application to the Superintendent of Police under Section 154(3), then to a magistrate under Section 156(3)
Direction to investigate	Section 156(3)	Magistrate empowered under Section 190	None prescribed	The complaint route remains available if the police still do not act	Complaint to the magistrate under Section 200
Arrest without warrant	Sections 41 and 41A	Police officer	Reasons recorded at the time of arrest	Arrest without the statutory satisfaction is unlawful and auditable by the magistrate	Notice of appearance instead of arrest; challenge at first production
Production after arrest	Sections 57 and 167	Police, then Judicial Magistrate	Within 24 hours, excluding journey time	Detention beyond the period without a magistrate’s order is unauthorised	Release, and habeas corpus in case of continued illegal custody
Remand during investigation	Section 167	Judicial Magistrate	Police custody only within the first 15 days	Custody beyond the authorised form or period is unlawful	Oppose extension at each remand hearing; seek bail
Default bail point	Section 167(2) proviso	Magistrate, on the accused’s application	90 days for offences punishable with death, life imprisonment or a minimum of ten years; 60 days otherwise	The right to bail accrues and is indefeasible if claimed before the charge sheet is filed	Application for default bail
Investigation in a summons case	Section 167(5)	Magistrate	Investigation to conclude within 6 months of arrest	Magistrate shall stop the investigation unless special reasons justify continuing	Application to stop the investigation
Police report	Section 173	Officer in charge of the police station	On completion of investigation, without unnecessary delay	Delay beyond the Section 167 period triggers the default bail consequence	Default bail; protest petition if a closure report is filed
Cognizance	Section 190	Magistrate	None prescribed	A closure report is not binding on the magistrate	Protest petition and hearing before acceptance of closure
Complaint case	Sections 200 to 203	Complainant, then Magistrate	Complainant examined on oath at presentation	Complaint may be dismissed under Section 203 if no sufficient ground appears	Inquiry or directed investigation under Section 202
Bail in non-bailable offences	Sections 437 and 439	Magistrate; High Court and Court of Session under Section 439	Heard as a judicial proceeding, no fixed period	Continued custody must be justified at each stage	Successive application to the higher forum
Anticipatory bail	Section 438	High Court or Court of Session	Available before arrest, on apprehension of arrest	Arrest without the protection if no application is made	Application under Section 438 before arrest
Maintenance	Section 125	Magistrate of the first class	Summary proceeding	Default in payment is enforceable by warrant and imprisonment as the section provides	Application for alteration of the allowance under Section 127
Judgment and beyond	Sections 353, 374 and 386	Trial court, then appellate courts	Appeal periods are prescribed by the law of limitation	Conviction or acquittal takes effect subject to appeal	Appeal, revision and the inherent powers under Section 482

Read down the limit column and the article’s claim is visible in data form. Nearly every stage at which the state holds power over a person’s liberty is attached to a number of hours, days or months, and nearly every number is attached to a consequence that operates automatically rather than by grace. That is the Code’s enforcement theory: not exhortation to the police or the magistracy, but defaults that fire when the powerful party misses its deadline. The theory does not always survive contact with a crowded cause list, which is the subject of the assessment that follows, but as a piece of institutional design it explains both the statute’s durability and the style of litigation it generates, in which the calendar is evidence and arithmetic is argument.

Replacement: The Bharatiya Nagarik Suraksha Sanhita and the Pending Cases

The Code’s long life ended as part of a package. In 2023 Parliament enacted three new codes to replace the three pillars of the colonial and postcolonial criminal law together: the Bharatiya Nyaya Sanhita for the substantive offences of the Penal Code, the Bharatiya Nagarik Suraksha Sanhita for the procedure of the 1973 Code, and the Bharatiya Sakshya Adhiniyam for the law of evidence. The Bharatiya Nagarik Suraksha Sanhita 2023 received assent in December 2023 and came into force on 1 July 2024, repealing the Code of Criminal Procedure 1973 for the future. The replacement was not a verdict that the 1973 Code had failed. Its architecture, its sequence of timed decision points, its classification grid and much of its language were carried across into the successor, renumbered and in places redesigned; what changed ownership was the text, not the underlying theory of the process. The successor’s scheme, its renumbering of the provisions this article has traced and its own substantive departures, are treated in the complete guide to the Bharatiya Nagarik Suraksha Sanhita 2023, which owns that statute’s account for this series.

What the repeal did not do is disturb the past. Section 531 of the Bharatiya Nagarik Suraksha Sanhita, its repeal and savings provision, provides in substance that any appeal, application, trial, inquiry or investigation pending immediately before the commencement of the new Sanhita shall be disposed of, continued, held or made in accordance with the Code of Criminal Procedure 1973 as in force immediately before that commencement, as if the new code had not come into force, subject to the specific exceptions the provision itself contains. The consequence is a two track present. Cases that began under the old Code run to their conclusion under the old Code, with its section numbers, its remand periods and its bail chapter, while cases beginning on or after 1 July 2024 run under the new one. Courts, police stations and law offices therefore work with both texts at once, and will for years, because the pending stock of Indian criminal litigation is measured in years of arrears. The 1973 Code is repealed the way a much litigated statute usually is: prospectively, completely, and with a long tail in which it remains the governing law for everyone whose case predates its successor. A reader who treats the Code as a closed historical subject will misread the reports; a reader who treats it as fully alive will misread the calendar. It is a repealed statute that continues to decide remand applications every working day.

For the practitioner moving between the two tracks, the working map between the texts is a short one at the points of highest traffic. The first information report that was Section 154 of the old Code is Section 173 of the Sanhita; the arrest and notice provisions of Sections 41 and 41A live in Section 35; the remand and default bail machinery of Section 167 is Section 187; anticipatory bail, Section 438, is Section 482 of the Sanhita; the inherent powers saved by Section 482 of the Code are saved by Section 528 of the Sanhita; the maintenance jurisdiction of Section 125 is Section 144 of the Sanhita, while the prohibitory orders that were Section 144 of the Code are Section 163 of the Sanhita, a collision of numbers that has already generated its share of confused citations. The full renumbering and the substantive changes that accompany it belong to the Sanhita’s own guide; what matters for the pending cases is the converse proposition, that in those cases none of these correspondences needs to be consulted at all, because the old numbers continue to govern by force of the savings provision.

What Works, What Does Not: An Honest Assessment

The case for the Code, stated at its strongest, is that it got the architecture right. It separated the trying magistracy from the prosecuting executive, which was the constitutional reform its century required. It organised the criminal process as a sequence of judicially supervised decision points with clocks attached, so that custody must be rejustified, investigation must conclude or release its grip, and closure must survive a hearing. It built safety valves at every level: the complaint route behind the police station, the protest petition behind the closure report, default bail behind investigative delay, anticipatory bail behind the abusive accusation, Section 482 behind the abusive proceeding. Its leading judicial interpretations have mostly developed, rather than contradicted, its text: Lalita Kumari enforced the registration duty the section already stated, Arnesh Kumar enforced the arrest discipline Parliament had already written, and Sushila Aggarwal read the anticipatory bail power at the width Sibbia had given it. A statute whose important cases consist largely of courts insisting that its text be obeyed is a statute whose text was, in the main, well judged.

The case against the Code’s working life, stated at its strongest, is that its disciplines are honoured unevenly at exactly the points where the design is most dependent on institutional behaviour. Arrest remained routine in practice long after the statute made it conditional, which is why a Supreme Court judgment was needed in 2014 to instruct the police in the contents of a 2008 amendment, and why the same Court has had to repeat the instruction since. Remand hearings in busy magistrates’ courts can shrink to a formality measured in seconds, at which the case diary is not read and the checklist is not audited, so that the judicial supervision the Code promises is supplied in appearance rather than in substance. The undertrial population, persons in prison awaiting the conclusion of their cases, has for decades formed a very large share of the prison population and frequently the majority of it, a pattern that sits badly with a code whose liberty defaults are as strong as this one’s; the figure moves from year to year and no single number should be attached to it, but the direction and the scale of the pattern are not in dispute. And the Code’s timelines, so precise for investigation, largely abandon the case at the trial stage, where no equivalent default fires when a trial itself stretches across years. Professor K N Chandrasekharan Pillai and other scholars of criminal procedure have long made the connected point that India reformed the law of investigation more thoroughly than the law of trial management, and the Code’s text bears that criticism out.

The most ambitious attempt to answer the Code’s critics from within the reform establishment was the report of the Committee on Reforms of the Criminal Justice System, chaired by Justice V S Malimath, submitted in 2003. The Malimath Committee argued that the system had tilted too far toward the accused and away from the victim and the truth, and recommended, among much else, a move toward features of the inquisitorial tradition, a lower standard of proof in the form of a middle path between proof beyond reasonable doubt and preponderance of probabilities described as the court’s conviction that the charge is true, and a dilution of the protections around silence and police statements. The recommendations were received as a manifesto by those who blame acquittal rates on procedural protection, and were opposed, by name and in print, by human rights lawyers and criminal law scholars who argued that the report mistook the cause of the system’s failures: that Indian trials fail on investigation quality, witness protection and court capacity, not on an excess of defence rights, and that weakening the standard of proof in a system with the undertrial record described above would multiply wrongful convictions among the poor. Parliament enacted neither the Malimath programme nor its opposite. The debate is worth recording in a pillar because it identifies the Code’s true contested frontier, which is not any single section but the standard of proof and the treatment of police evidence, the two load-bearing choices of 1973 that the Committee proposed to revisit and that every government since has declined to touch.

The verdict this pillar defends is that the Code of Criminal Procedure 1973 succeeded as a design and was defeated, where it was defeated, in its administration rather than its drafting. The distinction matters for how its replacement should be judged. If the failures were failures of text, a new code could cure them; if they are failures of behaviour, staffing and court capacity, the new code inherits them unchanged unless those conditions change, and the early experience under the successor will be the evidence. What can be said without waiting for that evidence is narrower and firmer. The 1973 Code converted a colonial administrative process into a constitutional procedure with a theory of liberty built into its machinery, and it did so durably enough that its successor kept the machinery. The general position described in this article is legal information and analysis, not advice on a specific case, and a person facing an actual proceeding should take advice on its facts, because the Code’s deadlines reward those who act on them in time and do not forgive delay.

That is also the note on which the article’s opening claim should be closed. The Code is a clock with two hands, and the hands still move in every pending case in the country: one counting the investigation toward the day the charge sheet must arrive or the accused walks out on default bail, the other counting the accused’s custody through remand hearings at which a magistrate must decide, again, whether the state may keep what it has taken. Whoever understands those two movements understands the Code, and understands why a statute of 484 sections can be summarised, without much loss, as a set of instructions about time.

Frequently Asked Questions

Q: What did the Code of Criminal Procedure 1973 replace?

The Code of Criminal Procedure 1973 replaced the Code of Criminal Procedure 1898, which had governed criminal process in India for three quarters of a century and was itself the last of a line of colonial codes enacted in 1861, 1872 and 1882. The replacement was comprehensive rather than partial: the 1973 Code, enacted as Act 2 of 1974 on the basis of the Law Commission’s Forty-first Report, repealed the 1898 Code and restated the whole field of criminal procedure, from the constitution of criminal courts to investigation, trial, bail and appeal. Much of the older machinery was carried across, including the classification of offences and the remand structure, but two changes were new in principle: the separation of the judiciary from the executive in the magistracy, and the creation of anticipatory bail, which the 1898 Code had never contained.

Q: Why did the Code of Criminal Procedure 1973 separate the judiciary from the executive?

The separation answered a structural defect of the older system, under which the magistrates who tried criminal cases were subordinate to the District Magistrate, the head of the district executive, so that the officer deciding a prosecution sat inside the same administrative hierarchy as the police who had investigated it and the government that employed him. Even where no pressure was applied, the arrangement compromised the appearance and the institutional reality of independent adjudication. The Constitution’s directive principle in Article 50 directed the state to separate the judiciary from the executive in the public services, and the Law Commission’s Forty-first Report recommended carrying that direction into the district criminal courts. The 1973 Code performed it by placing Judicial Magistrates, who try cases, authorise custody and decide bail, under the High Court, while leaving Executive Magistrates, who handle preventive and administrative functions, under the state government.

Q: How does default bail work when a charge sheet is filed late?

Default bail arises under the proviso to Section 167(2). If the investigation is not completed and the police report is not filed within ninety days where the offence is punishable with death, life imprisonment or imprisonment of not less than ten years, or within sixty days in any other case, the accused acquires a right to be released on bail on the expiry of the period, provided he applies for it and is prepared to furnish bail. The Supreme Court has described the right as indefeasible: the magistrate cannot refuse it because the accusation is serious. The timing of the charge sheet is therefore decisive. If the accused claims the right after the period expires and before the charge sheet is filed, the later filing cannot defeat it; if the charge sheet is filed first and the accused has not yet claimed the right, the right is extinguished and bail must be sought on the merits of the case.

Q: Where did anticipatory bail come from?

Anticipatory bail was created by Section 438 of the Code of Criminal Procedure 1973 and had no equivalent in the Code of 1898, which offered protection only after arrest through the ordinary bail provisions. Its source is the Law Commission’s Forty-first Report, which recommended the jurisdiction because the Commission had observed a recurring abuse: persons were implicated in false criminal cases, sometimes out of rivalry, enmity or political motive, in order to secure their arrest and the disgrace and detention that arrest involved, even though the accusation later collapsed. A remedy that began only after arrest could not answer an abuse whose whole point was the arrest itself. Parliament accepted the recommendation, empowering the High Court and the Court of Session to direct that a person who has reason to believe he may be arrested on an accusation of a non-bailable offence shall, in the event of arrest, be released on bail on conditions the court imposes.

Q: What are the inherent powers of a High Court under the Code of Criminal Procedure?

Section 482 of the Code saves the inherent powers of the High Court, providing that nothing in the Code limits the High Court’s power to make such orders as may be necessary to give effect to any order under the Code, to prevent abuse of the process of any court, or otherwise to secure the ends of justice. The section does not create the powers; it recognises powers the High Court possesses independently and preserves them against the argument that the Code’s express provisions exhaust its authority. In practice the jurisdiction is used most often to seek quashing of a first information report or of criminal proceedings, and the Supreme Court in State of Haryana v Bhajan Lal set out the categories in which quashing is proper, such as allegations that disclose no offence even if accepted entirely. The power is exercised sparingly, and the High Court does not weigh evidence or conduct a trial within the petition.

Q: What did Lalita Kumari require about registering an FIR?

The Constitution Bench in Lalita Kumari v Government of Uttar Pradesh, decided in November 2013, held that registration of a first information report under Section 154 is mandatory where the information given to the police discloses the commission of a cognizable offence. The officer in charge has no discretion to test the truth of the allegation before registering it, because verification belongs to the investigation that follows registration. The Bench allowed a preliminary inquiry in a limited class of cases, naming matrimonial and family disputes, commercial offences, medical negligence cases, corruption cases and cases of abnormal delay in launching prosecution as examples, but confined the inquiry to whether a cognizable offence is disclosed at all, required it to be time bound, and barred its use as a general screening of complaints. The effect is that a police station may not refuse registration because the allegation seems improbable, inconvenient or unlikely to end in conviction.

Q: What were the Arnesh Kumar guidelines on arrest?

In Arnesh Kumar v State of Bihar, decided in 2014, the Supreme Court addressed routine arrest in offences punishable with imprisonment of up to seven years, the class governed by the checklist in Section 41 of the Code. The Court directed that police officers must not arrest automatically on the registration of such a case but must satisfy themselves, against the sub-clauses of Section 41(1)(b)(ii), that arrest is necessary for purposes such as proper investigation, preventing further offences, or preventing tampering with evidence or flight, and must record their reasons. Officers were to be provided with the checklist itself. Magistrates authorising detention were directed not to act casually or mechanically but to record their own satisfaction that the arrest was justified, and a police decision not to arrest was to be forwarded to the magistrate within the period the judgment specified. The Court warned that unjustified arrests and mechanical remands could attract departmental action and contempt consequences.

Q: Is the Code of Criminal Procedure still used in pending cases?

Yes. Although the Bharatiya Nagarik Suraksha Sanhita 2023 repealed the Code of Criminal Procedure 1973 and replaced it for cases beginning on or after 1 July 2024, the savings provision in Section 531 of the Sanhita directs that any appeal, application, trial, inquiry or investigation pending immediately before that commencement shall be disposed of, continued, held or made in accordance with the Code of Criminal Procedure 1973 as it stood before the repeal, subject to the exceptions the provision itself states. Pending cases therefore continue under the old Code’s section numbers, remand periods, default bail rules and bail provisions until they conclude, while fresh cases proceed under the Sanhita. Because the pending stock before Indian criminal courts is large and moves slowly, the 1973 Code remains the governing procedural law in a substantial part of the criminal litigation in the courts and is applied in remand and bail hearings every working day.

Q: What is the difference between a cognizable and a non-cognizable offence?

The difference is the power of the police to act on their own authority. In a cognizable offence, a police officer may arrest the accused without a warrant and may begin an investigation without any order from a magistrate, because the offences classified as cognizable are generally the more serious ones in the First Schedule. In a non-cognizable offence, the officer has neither power: no arrest may be made without a warrant, and no investigation may be undertaken except under the order of a magistrate, as Section 155 provides. The classification is fixed offence by offence in the First Schedule to the Code for offences under the Indian Penal Code and by the governing statute for offences under other laws. The distinction decides how a case can begin at the police station, whether the first information report route under Section 154 is available at all, and whether the person accused can be taken into custody on the spot or only after a court has been moved.

Q: What remedies does the Code give when the police refuse to register an FIR?

The Code builds a ladder of remedies that does not depend on the refusing officer changing his mind. First, under Section 154(3), the substance of the information may be sent in writing to the Superintendent of Police, who, if satisfied that the information discloses a cognizable offence, must investigate the case himself or direct a subordinate officer to investigate it. If that fails, the informant may apply to a magistrate under Section 156(3), and the magistrate may order the police to register and investigate the case, a power the Supreme Court in Sakiri Vasu held includes the ability to monitor the investigation ordered. The informant may alternatively bypass the police altogether and file a complaint before a magistrate under Section 200, whereupon the magistrate examines the complainant on oath and proceeds under the complaint procedure. The writ jurisdiction of the High Court stands behind these remedies, but the Court in Sakiri Vasu directed that the Code’s own ladder should ordinarily be used first, because it is efficacious.

Q: What is the difference between police custody and judicial custody during remand?

Both are forms of detention authorised by a magistrate under Section 167 during investigation, but they differ in who holds the accused and what may be done with him. In police custody the accused remains with the investigating agency and is available for interrogation and for investigative steps such as recoveries and identification, and the Code confines police custody to the first fifteen days after the accused is produced before the magistrate. In judicial custody the accused is lodged in prison under the authority of the court; the police may not interrogate him as a matter of routine and must work through the court for any access. The distinction is a deliberate design choice: intensive police access is permitted briefly at the start of the investigation, when evidence is fresh, and is then withdrawn, so that the longer part of any pre-trial detention is spent in neutral custody where the pressure of interrogation does not operate. A magistrate chooses the form of custody at each remand hearing and must record the authorisation as a judicial act.

Q: Why are some statements recorded before a magistrate under Section 164?

Because statements recorded by the police during investigation are deliberately given a weak evidentiary status, and Section 164 provides the guarded channel through which a statement can acquire a stronger one. A statement to the police under Section 161 cannot be signed by its maker and cannot be used as evidence of its contents at trial, except to contradict the witness. A confession or statement recorded by a Judicial Magistrate under Section 164, by contrast, is recorded under safeguards designed to ensure it is voluntary: the magistrate must warn the person that he is not bound to confess and that a confession may be used against him, must allow time for reflection away from police custody, and must certify that the statement was made voluntarily. A confession recorded through this channel can be used at the trial, which is precisely why the safeguards are enforced strictly and why a confession obtained without them is liable to be rejected. The channel also serves witnesses, whose statements a magistrate may record to preserve their evidence in authoritative form.

Q: What can the informant do if the police file a closure report instead of a charge sheet?

The informant is not bound by the police conclusion, and the case is not over when a closure report reaches the court. Under the rule settled in Bhagwant Singh v Commissioner of Police, the magistrate must give the informant notice and an opportunity of being heard before accepting a final report that closes the case, because the informant set the criminal law in motion and has a recognised interest in its outcome. At that hearing the informant may file a protest petition disputing the closure and pointing to the material the investigation gathered or ignored. The magistrate then has real choices: he may accept the closure notwithstanding the protest, he may treat the protest petition as a complaint and proceed under Sections 200 and the following provisions, examining the informant on oath, or he may disagree with the police and take cognizance of the offence on the material before him under Section 190. The closure report is therefore a recommendation to the court, not a verdict, and the informant’s hearing is the mechanism that keeps it so.

Q: What is the difference between bail in a bailable offence and bail in a non-bailable offence?

In a bailable offence, bail is a right. Section 436 provides that a person arrested for a bailable offence shall be released on bail if he is prepared to give bail, and neither the police officer at the station nor the court has a discretion to refuse on the merits of the accusation; the classification in the First Schedule has already decided that custody is not ordinarily warranted. In a non-bailable offence, bail is a matter of judicial discretion under Section 437 in the magistrate’s court and Section 439 in the High Court and the Court of Session, exercised by weighing the gravity of the charge, the severity of the punishment, the risk of flight, the danger of tampering with evidence or influencing witnesses, and the character of the material. The distinction is not between release and detention as outcomes, since most accused in non-bailable cases are also released, but between a claim the accused can enforce by right at the station and an application he must argue for before a court.

Q: What does it mean for a magistrate to take cognizance of an offence?

Cognizance is the magistrate’s formal act of applying his mind to the facts before him, as disclosing the suspected commission of an offence, with a view to initiating judicial proceedings. Section 190 permits it in three situations: upon a complaint, upon a police report, and upon information from any other person or the magistrate’s own knowledge. Two clarifications from the case law govern its use. Cognizance is taken of the offence and not of the offender, so the act concerns whether a triable case appears to exist rather than whether a particular accused is guilty. And taking cognizance is not a finding on the merits; it is the decision that moves the case from the investigative or complaint stage into the trial process, after which the accused is summoned, the charges are framed in due course, and the evidence is tested. A magistrate may also decline to take cognizance, most visibly by accepting a closure report after hearing the informant or by dismissing a complaint under Section 203, which is why the stage functions as a judicial filter between accusation and trial.

Q: Can a private person start a criminal case by complaining directly to a magistrate?

Yes. The complaint route in Sections 200 to 203 is the Code’s standing alternative to a police case, open to any person, and it does not require the police to act at all. The complainant presents the complaint to a magistrate competent to take cognizance; the magistrate examines the complainant and any witnesses present on oath and records the substance of the examination. He may then hold an inquiry himself or direct an investigation by the police or another person under Section 202, postponing the issue of process meanwhile. If he finds no sufficient ground for proceeding, he must dismiss the complaint under Section 203 and briefly record his reasons; if he finds sufficient ground, he issues process to the accused under Section 204 and the case proceeds as a complaint case through the trial chapters. The oath, the inquiry and the reasoned dismissal power are the filters that keep the route from becoming an instrument of harassment, but within those filters the route is complete: it can carry a case from accusation to judgment without a police station ever being involved.

Q: What is a Section 144 order and how long can it remain in force?

A Section 144 order is a written prohibitory order by which a District Magistrate, a Sub-divisional Magistrate or an empowered Executive Magistrate, in an urgent case of nuisance or apprehended danger, directs persons to abstain from a specified act or to take specified order with certain property, stating the material facts of the situation it answers. It is the Code’s standard instrument for preventing assemblies, processions and other conduct thought likely to cause a disturbance of public tranquillity, and it operates in anticipation of danger rather than in punishment of an offence. Its defining legal feature is brevity. An order cannot remain in force for more than two months from its making; the state government may extend it by notification where necessary to prevent danger to human life, health or safety or a likelihood of riot or affray, but the total duration cannot exceed six months. A person aggrieved may apply to the magistrate for rescission or alteration and is entitled to an early hearing, which is the route by which overlong or formulaic orders are tested.

Q: What is the difference between investigation, inquiry and trial under the Code?

The Code defines the three stages separately because different actors and powers belong to each. Investigation, defined in Section 2(h), consists of all proceedings under the Code for the collection of evidence, conducted by a police officer or by a person authorised by a magistrate; it produces the material, the statements, the seizures and the police report, but decides nothing. Inquiry, defined in Section 2(g), is every inquiry other than a trial conducted under the Code by a magistrate or court, such as the preliminary examination of a complaint or committal type proceedings; it is judicial, but it does not determine guilt. Trial is the stage the Code uses without a comprehensive definition, in which the court hears the evidence for the prosecution and the defence, frames and tests the charge, and delivers judgment of conviction or acquittal. The sequence matters practically: powers of arrest, search and recording attach to investigation, the filter of cognizance and dismissal attaches to inquiry, and the full protections of the presumption of innocence, proof beyond reasonable doubt and cross-examination attach to the trial.

Q: What is plea bargaining and when was it added to the Code?

Plea bargaining was added to the Code by the Code of Criminal Procedure (Amendment) Act 2005, which inserted Chapter XXI-A, comprising Sections 265A to 265L, as the legislature’s response to the accumulation of pending trials. The scheme is narrower than its counterparts in some other systems. It is available to an accused charged with an offence not punishable with death, imprisonment for life or imprisonment exceeding seven years, and it excludes offences against women and children and the socio-economic offences notified for the purpose. The accused applies to the court trying the case; the court examines the application and the voluntariness of the choice, and a mutually satisfactory disposition is then worked out with the participation of the public prosecutor, the investigating officer where the case began on a police report, the victim and the accused, with compensation to the victim forming part of the disposition. The court passes judgment in accordance with the disposition, awarding the agreed sentence which the statute moderates, and the judgment is largely final. The design trades a measure of punishment for certainty, speed and compensation, within judicial supervision.

Q: How does the victim compensation scheme under Section 357A work?

Section 357A, inserted by the amendment of 2008 with effect from 31 December 2009, requires every state government, in coordination with the Central Government, to prepare a scheme providing funds for compensation to a victim or the victim’s dependents who have suffered loss or injury as a result of a crime and who require rehabilitation. The scheme is administered through the Legal Services Authorities. A court may recommend compensation under the scheme where the compensation otherwise payable is inadequate for rehabilitation, or where the case ends in acquittal or discharge and the victim still requires rehabilitation; where the offender is not traced or identified and no trial takes place, the victim or dependents may apply directly to the State or District Legal Services Authority, which must complete its enquiry and award adequate compensation within the period the section prescribes. The Authority may also order immediate first aid and free medical benefits as interim relief. The design separates compensation from the fate of the prosecution: it is a welfare entitlement administered by a legal services institution, funded by the state, and available even where no conviction and in some situations no identified offender exists.