It is ordered by the court that the order of this court entered in this case on March 9, 1959, denying the petition for leave to prosecute an appeal in forma pauperis is vacated.

It is further ordered by the court that petitioner is allowed to proceed on appeal from the judgment of conviction without prepayment of costs and that the joint appendix shall be printed at the expense of the United States.

Each judge reserves the right to file a statement of his position in this matter.

WILBUR K. MILLER, Circuit Judge, dissents.

Statement of Circuit Judge BAZELON: This case is before us on a petition for leave to prosecute in forma pauperis an appeal from a judgment of conviction under the narcotics laws, 26 U.S.C.A. 4701 et seq.1 On March 9, 1959, by an order from which I dissented, this court denied the petition. It was noted in the order that each judge reserved the right to file a statement of his position. Since then, by order of April 30, 1959, the court reversed itself and granted the petition, Circuit Judge Miller dissenting. I deem it advisable to state my reasons for voting to grant the petition.

We are required to grant the petition if the appeal presents a question which is not plainly frivolous. Ellis v. United States, 1958, 356 U.S. 674, 78 S. Ct. 974, 2 L. Ed. 2d 1060. The Government opposes this petition on the ground that the question presented is plainly frivolous. The Government almost invariably opposes such petitions on this ground.2 Ordinarily, however, we appoint counsel to file a memorandum in support of the petition, and we grant most of the petitions.3 3] In a majority of these appeals in forma pauperis we affirm-- as we do in most 'paid' appeals.4 But there are many meritorious appeals among them, and a substantial number result in reversals. If, because of the Supreme Court's Ellis decision, we are allowing more criminal appeals in forma pauperis, it is probable that these include a larger number of meritorious appeals.

From September 1, 1957, through February 28, 1959, we decided 24 appeals in which we had granted leave to appeal in forma pauperis although the District Court had previously denied leave on the ground that the appeal lacked merit.5 We reversed the convictions in 11 of these 24 cases.6 In 6 of the other cases the convictions were affirmed by split decisions.7 Thus in 17 of the 24 cases at least one judge of this court thought the conviction should be reversed. In all 17 of these cases, the Government had opposed the petitions for leave to appeal in forma pauperis.

A free society should be ready to assume the burden of infinitely more than 24 appeals to avert 11 miscarriages of justice. Concern for our docket should not attenuate the right of every indigent to point out, with the aid of counsel, what may prove to be errors. If time of bench and bar is spent in review of cases lacking in merit, it is the price we must pay for a system of justice that shields the poor man from 'invidious discriminations.'8 We cannot review the cases which prove to have merit without reviewing some which may not.

I do not minimize the heavy burdens which the bar is called upon to bear in representing indigent appellants in this court.9 We hope that Congress will provide an adequate solution. Until it does, in predicting which appeals may prove valid, our concern for the worth and dignity of every individual dictates that we had best err on the side of indulgence.

The question for which review is being sought in the present case is whether, in testing the sufficiency of cause for an arrest based on an informer's report, the accused is entitled to know the informer's identity.

The record discloses the following: On March 6, 1958, Officer Ernst of the Metropolitan Police Department Narcotics Squad received a telephone call from a confidential informant in New York City to the effect that petitioner had purchased narcotics in New York and was returning to Washington by train. The informer had supplied information in the past which the officer had found to be reliable. Officer Ernst, who did not know petitioner by sight, conveyed the information to Officer Somerville, who did know petitioner by sight. Officer Somerville met the train and, acting solely on the report conveyed to him by Officer Ernst,10 arrested petitioner. He made an immediate search of petitioner's person and found certain narcotics in his pocket.

A pre-trial motion to suppress the evidence was denied. At the hearing of the motion, the prosecutor invoked the governmental privilege not to disclose the identity of the informer. The court sustained the privilege and defense counsel acquiesced in the ruling. At trial, defense counsel again raised the question of petitioner's right to disclosure of the informer's identity. The court ruled that the question was foreclosed by the earlier ruling sustaining the Government's privilege.

The narcotics were received in evidence at trial and provided the basis for petitioner's conviction on two counts for which he was given concurrent sentences of two to seven years, and seven years.

The exigencies of law enforcement require that a police officer be permitted to make an arrest when he has probable cause to believe a person guilty of a felony. Ordinarily he must present what he knows to a magistrate on an application for a warrant, so that the decision whether or not to arrest may be made as judiciously as possible. But, when circumstances preclude this judicious procedure, the police officer may make his own determination of probable cause, subject to later judicial testing. This represents a compromise with principle, based on the assumption, not always correct,11 that when the objectivity of a magistrate cannot be brought to bear, a police officer sworn to uphold the law may be relied upon to act judiciously in abridging an individual's liberty. This compromise, which is dictated by necessity, transfers from the judicial officer to the police officer the delicate judicial function of determining, in the first instance, whether there is probable cause. The dangers are plain. They would be increased beyond limits tolerable in a free society if the law allowed the police officer to arrest solely on the basis of an informer's uncorroborated assertions. For this would give the informer the delicate function which the law generally reserves for the judicial officer and allows the police officer to exercise only in limited situations. The considerations which justify delegation of the judicial function to police officers do not justify a further delegation to their informants.

It is notorious that the narcotics informer is often himself involved in the narcotics traffic and is often paid for his information in cash, narcotics, immunity from prosecution, or lenient punishment. In Brandon v. United States, No. 14464, now pending before us, the informer was a narcotics user who was out on bail on another narcotics charge when she gave her information. Members of the very Narcotics Squad to which Officers Ernst and Somerville belong have in the past supplied drugs to an informer-addict who had difficulty obtaining supplies through normal illicit channels.12 The so-called 'special agent' or 'special employee' of the Metropolitan Police Department who served as informer in Smith v. United States, 1958, 103 U.S.App.D.C. 48, 254 F.2d 751, was then under an indictment and had had numerous convictions including one for a narcotics violation. He was paid no salary but received a fee for his information.13

For such reasons, the law has wisely circumscribed the use of informers' reports as a basis for making arrests. It has required a showing of reliability of the information17 and some corroboration by facts within the arresting officer's own knowledge. Such corroboration may be obtained by placing the suspect under surveillance and observing whether his behavior-- e.g., making contact with persons known to be in the trade, or surreptitious passing of a package-- tends to support the informer's story.

How much personal knowledge the policeman must have before making the arrest cannot be determined for all cases. Since circumstances vary, the rule must be stated in terms of what is reasonably inferable from the particular circumstances by a prudent police officer.18 In the present case, the informer had reported that petitioner would be getting off a train from New York with narcotics. That the officer had sufficient basis for making the arrest when he saw the petitioner get off the train is settled by the Supreme Court's recent decision in Draper v. United States, 1959, 358 U.S. 307, 79 S. Ct. 329, 3 L. Ed. 2d 327. It would appear from Draper that when a reliable informer tells an officer that an individual has left town to buy narcotics and is returning by train at a certain time, the policeman's observation of his return is enough to justify an arrest without a warrant.

When deprivation of liberty by arrest is justified by a combination of an informer's report and a policeman's own observations, the less the policeman is required to observe, the more significant becomes the informer's report. Since observation of so commonplace an act as coming home by train is enough, in combination bination with the report of an informer, to justify an arrest, the only significant limitation of the informer's power to cause an arrest is the requirement of reliability. The requirement that the informer be reliable stands as the only effective legal safeguard against false denunciations by irresponsible individuals who may be motivated by self-interest, spite, or even paranoia. The only other safeguard which remains rests not on law but on the good will of the police officer.

To give vitality to the requirement of reliability the law requires, if no probable cause exists without the information from the informer, that his identity be divulged, so that his reliability may be subjected to meaningful judicial scrutiny rather than accepted on a policeman's word. Roviaro v. United States, 1957, 353 U.S. 53, 77 S. Ct. 623, 1 L. Ed. 2d 639. In holding in Roviaro that the Government's 'informer privilege' must give way to the accused's need for disclosure, the Court was dealing with the problem in a context other than the legality of an arrest, but the principle clearly extends to this context. The Court said:

'Where the disclosure of an informer's identity, or the contents of his communication, is relevant and helpful to the defense of an accused, or is essential to a fair determination of a cause, the privilege must give way. * * * Most of the federal cases involving this limitation on the scope of the informer's privilege have arisen where the legality of a search without a warrant is in issue and the communications of an informer are claimed to establish probable cause. In these cases the Government has been required to disclose the identity of the informant unless there was sufficient evidence apart from his confidential communication.' Id., 353 U.S. at pages 60-61, 77 S.Ct. at page 628.19

It is easy to argue that the informer's reliability is established by the fact that the petitioner really did have narcotics in his pocket when he was arrested. But it is too well established for argument that the validity of an arrest cannot be proven by its fruits.20 A police officer may not make an arrest on the basis of information supplied by an informer unless he has reason to believe that on a post-arrest judicial test the informer's report can be shown to have been reliable when the officer acted on it. This is not a technical rule designed to shield the relatively few persons who are ultimately charged with crime; it is designed for the protection of the many whom the police arrest but do not charge. There is reason to believe that for every illegal arrest that produces evidence leading to prosecution, there are many which produce nothing, so that no charges are made and the arrests never come to light.21 The many innocent persons thus arrested have no effective recourse against anyone for the deprivation of their liberty.22 The law recognizes that the only way to protect the innocent is by imposing safeguards which protect the guilty as well. Weeks v. United States, 1914, 232 U.S. 383, 34 S. Ct. 341, 58 L. Ed. 652.

Under the principle of Roviaro, the District Court could not properly determine that there was probable cause for petitioner's arrest without requiring disclosure of the identity of the informer whose report was claimed to constitute such cause. Nothing in Draper suggests that the Court has abandoned the Roviaro principle. Indeed the Court pointed out in Draper that the identity of the informer was disclosed. The treatment of the requirement of corroboration in Draper suggests that the Roviaro principle must be strictly enforced.

At the very least, it seems to me, there is no basis for concluding that the issue here presented is plainly frivolous. Ellis v. United States, 1958, 356 U.S. 674, 78 S. Ct. 974, 2 L. Ed. 2d 1060.