with whom Circuit Judges GEWIN, BELL, THORNBERRY, COLEMAN, AINS-WORTH, SIMPSON, and CLAYTON join, concurring specially:
I concur fully in the Court’s opinion and the result. Some comments in the dissenting opinion warrant some emphasis.
The dissenting opinion deplores what it senses is undue emphasis by the Court on the so-called technicalities of perjury, presumably Texas peculiarities, even though its concepts seem to be those everywhere held. But for an otherwise valid state conviction to be upset years later on federal habeas, surely something more than an evidentiary 'mistake must be shown. If mistake is enough, then never, simply never, will the process of repeated, prolonged, postconvietion review cease. For in every trial, or at least nearly every trial, there will be, there are bound to be, some mistakes.1
*41What elevates the “mistake” to a constitutional plane is at least twofold. First, the mistake must be material in the sense of a crucial, critical, highly significant factor. Second, it must have some State complicity in it. As to this, I do not for a moment think innocence of the prosecutor is an assured out. For Barbee properly teaches that the police are very much a part of the State and its prosecutorial machinery. But someone in the state machinery must have some awareness that testimony being palmed off as the gospel is something else indeed.
On these scores there is nothing to the assertion of “perjury” or “falsity.” First, the interrogator chose to east his questions in conclusory terms. Phrasing the question first, “Q. How many cases have they filed on you * * *?”, the examiner almost instantaneously pinpointed it to the instant present— “* * * do they have over you right now?” To this conglomerate of conclusions, the witness answered in like terms: “A. I don’t have any pending cases against me.” It turns out, of course, that under Texas law, for all practical purposes Gober had no “pending” eases.2 So far as this formal charge was concerned, Gober was as free as the birds. While it may be true that new charges could be filed, or an indictment returned within the statute of limitations period, none had been, and to this date, none ever has been. One simply cannot make this negative reply a falsity merely because, had the questions been framed differently or the matter pursued to an obvious end for a counsel bent on questioning or hopefully destroying credibility, the witness would have had to state that new charges could be filed or an indictment returned. By the dissent perjury is first watered down to falsity, and then both are brushed aside by a statement of what all of us as Judges know concerning the possibility of new charges or indictments but as to which the witness was never interrogated. Worse, the only way to reach the conclusion that the witness was being interrogated along these shadowy, vague, unidentifiable lines is to make the triple assumption that (1) the interrogator had that in mind, (2) the witness knew the interrogator had that in mind, and (3) the witness by no inadvertent slip gave a known wrong answer.
Second, there is nothing to implicate the State in this in any way. The prosecutor in the courtroom was ignorant of the original charges. The police officers who would have had knowledge were out of the courtroom under the Buie.
On a like approach the charge that the “State knowingly suppressed evidence beneficial to the defendant Luna” or, stated more advantageously to the petitioner, the State failed to furnish information helpful to the defendent, turns out equally to be lacking in constitutional substance.
The assertion here is that in the testimony of Gober, the State misled the defendant because the testimony did not reflect the connection between Gober, as an informant, and the State. Emphasis is placed on his negative answer to the question “Are you paid by them [the Houston Police Department] ?”
Here again the answers stated in terms of the present were literally true. He had no “official capacity with the Houston Police Department,” he was not “a police officer”, and he was not “paid by them” in the sense of being a salaried employee. On the other hand, it was perfectly clear that he had been “working with or for the Houston Police De*42partment.” Indeed, Gober went further and pinpointed the time by stating that he had not worked with the Houston Police Department “since some time in July, the middle of July.”
From the circumstances of the sale of narcotics by Luna to Gober, Gober’s familiar role in the “making” of the sale and purchase and Gober’s accurate answers to the questions as put to him, it was perfectly evident to anyone having any experience that Gober obviously was being used by narcotics officers 3 in the enforcement of these important laws. No matter how much one might deplore the necessity for the use of informants and who up the limits of entrapment “stage” cases for a successful prosecution, this is both a well known and judicially acceptable phenomenon. And except for those possible and rare instances in which the detested role of informer is taken on out of a penitent conversion, the practice rests upon the advantages which the informer gets, or thinks he gets, at the hands of the sovereign. It may be, and often is, money — not only little handouts, but, as this Court has expressly approved, a contingent fee arrangement in which even the National sovereign is allowed to pay bounty on a sliding scale depending on the number of convictions the “special employee” brings about.4 In others, the advantage can be an indulgence either by police officers or, for that matter, persons as elevated in the hierarchy as the Attorney General of the United States5 toward the prosecution or even the commencement of criminal proceedings against the informant for his known criminal actions. In others it takes a middle course in which on arraignment and acceptance of a co-defendant’s plea of guilty the Court acquiesces in the prosecutor’s recommendation for a dismissal of most charges, reduction in the offense, and postponement of sentence with its telltale message that punishment may well depend on how well the penitent performs and the prosecution succeeds.
In the light of this record, there was a vivid revelation of Gober’s relation to the State in the prosecution and trial of this case (see note 3 supra). Once we assume that Gideon’s 6 trumpet heralds the importance of counsel because of professional knowledge and skill beyond the experience of the non-lawyer, no one worthy of the Bar could have claimed even remotely any surprise as to Gober's connection with the case, the role he was playing and the whole gamut of influences — monetary, physical, emotional, or psychological — which might be at work , in his helping, if not instigating, the efforts of the police. Spread out over the whole record for all to see and the skilled lawyer to pursue was Gober’s relationship. Unless, even though counsel is present and performing, the Federal habeas later look-back has to assume professional ignorance bordering on incompetence and an unawareness of what is going on bordering on mental unconsciousness, the defendant through his counsel knew all that the dissenters insist the State should have revealed. Perhaps it was a mistake for counsel not to have pursued it. But the mistake was not *43due to being mislead by appearances or ignorance of the real situation. The mistake — and as a tactical judgment of counsel in the heat of trial it may well not have been a mistake at all — was in counsel not following up the obvious.
The opportunity for full exploration was there. The certainty that Gober was acting out of supposed self-interest was patent. The failure to press on, the failure to pursue relentlessly was not from a State sponsored, State instigated or State permitted deficiency. Rather it was from a decision — deliberate or unconscious — by counsel. But that cannot make the trial unfair or afford to this defendant release in the high name of the constitution.
. Consider, for example, a defense of alibi refuted factually by a witness positively identifying the defendant and then rejected by the jury’s verdict of guilty. Five years later post-eonvietion proof demonstrates conclusively that the identification was erroneous although maintained by the witness, the police, and the prosecution in all good faith as the truth. This would, of course demonstrate that the result was wrong. But it would not demonstrate that the trial was unfair. The Federal Constitution could not correct that mistake.
. This point on tlio effect of expiration of the court term without the return of an indictment was advanced for the first time in the forcefully able brief and argument of the Assistant District Attorney to the Court en banc. Lacking this guidance, the initial panel was, of course, entitled to assume that the formal charge was viable.
. See the recitation of facts, Luna v. State of Texas, Tex.Crim.App., 1965, 387 S.W.2d 896, set out at length in the dissenting opinion: Gober calls officer Hightower. Gober played it easy to get time to contact officer Hightower at the Pigstand drive in, where Hightower, first searching Gober, furnished $21.00 to Gober. At the exchange of money for capsules Hightower and officer McMannes had Gober under continuous surveillance followed by search of Gober and delivery of narcotics to the officers.
. Williamson v. United States, 5 Cir., 1962, 311 F.2d 441, affirmed after retrial, 340 F.2d 612; Bullock v. United States, 5 Cir., 1967, 383 F.2d 545; Hill v. United States, 5 Cir., 1964, 328 F.2d 988.
. United States v. Cox, 5 Cir., 1965, 342 F.2d 167, (en banc) cert. denied sub nom. Cox v. Hauberg, 381 U.S. 935, 85 S.Ct. 1767, 14 L.Ed.2d 700.
. Gideon v. Wainwright, 1963, 372 U.S. 335, 83 S.Ct. 792, 9 L.Ed.2d 799.