Defendants appeal from convictions of possession of 30 pounds of marijuana in violation of 21 U.S.C. § 841(a)(1). The issue is the validity of the arrests of the defendants and the searches of two automobiles, each occupied by one of the defendants. We hold that the arrests and searches were invalid and reverse the convictions.

November 10, 1975, Agent Lofstrum of the Drug Enforcement Administration (DEA), presumably stationed at Brownsville, Texas, received from his supervisor an unsigned letter containing the following information:

Heroin in large amounts packed into spare tires in Elsa and then taken to Houston by organization Rene Villarreal ex convict, West 6th, Elsa, and Manuel Montalvo, pack spare tires full of heroin every few days.

Villarreal drives red Nova FST-807 and takes pregnant wife and child across Checkpoint at Sarita. Montalvo with other spare full of heroin meets him at road side park north of Rivera (right side). Montalvo drives brown Marquis FRA-470. Others with spare full of heroin follow 1 mile apart.

La Chiva red 67-FSM-466 carries spare. All spares are loaded at roadside park in one car and usually 4 spares of heroin go to Houston in one car from roadside north of Rivera. Cars travel through checkpoint at Sarita 1 mile apart.

This letter was addressed to "Dept. of Justice, Drug Enforcement, . . . Brownsville, Tex.," and was postmarked November 7, 1975, p. m., Combes, Texas.

The same day Lofstrum placed at the Sarita, Texas, permanent checkpoint a "lookout notice" for three vehicles as described in the letter. The notice gave the name Rene Villarreal and described the suspected modus operandi described in the letter.

Five or six days after he received the first letter, a second letter came into Lofstrum's hands. It was addressed to an employee of a Harlingen, Texas, television station who had referred it to the DEA. Stationery, envelope, and postmark were identical to the first letter. This letter was signed "Observer, Observer," and read as follows:

Please pass the following information to the right places.

Heroin in large amounts is being repacked in Elsa, Tex. for shipment to Houston. Rene Villarreal ex convict West 6 street, Elsa and Manuel Montalvo, Elsa pack 3 or more spare tires full of heroin every few days and take them to Houston. Villarreal drives a red Nova FST-807 and takes pregnant wife and child to Roadside park near Rivera. He meets Montalvo in brown Mercury Marquis FRA-470 and two or more other cars. Each car has a spare full of heroin and after check station cars meet and Villarreal loads spare tires full of heroin into one car usually brown Marquis FRA-470 and then goes to Houston.

We do not know whether the letters were written by the same or different persons. After receiving the second letter, Lofstrum renewed the lookout notice at the checkpoint on November 17.

On November 19, Charles McClure, a Border Patrol Officer, was working at the Sarita checkpoint. He was aware of the lookout notice posted by Lofstrum, and he knew of the information contained in the letters. At approximately 11:20 p. m. McClure observed, and stopped at the checkpoint, a 1970 brown Mercury Marquis with license FRA-470. The driver, defendant Smith, had no identification except an electric bill containing that name. One to two minutes after the arrival of the Mercury a 1974 red Chevrolet Nova two-door arrived. Its license number was FST-807. The driver was Villarreal. He produced identification matching the name given in the letters. With him were his pregnant wife and small child as described in the first letter. McClure took the cars and occupants into custody and notified DEA.

DEA Agent Lofstrum arrived at the checkpoint at approximately 2:00 a. m. November 20, and took custody of Smith and Villarreal and the two vehicles. Without securing a warrant he examined the spare tire in the trunk of each vehicle and by shaking them detected something loose inside. When he let air out of each one he could smell the odor of marijuana. He deflated and opened the tires and found brick marijuana, approximately 15 pounds in one tire and 13 in the other.

We hold there was no probable cause to arrest the defendants and search their cars.1

The tips in this case came from one or two unknown and anonymous informers. Our guidelines for determining whether there is probable cause based on an informer's tip come from Spinelli v. U. S., 393 U.S. 410, 89 S. Ct. 584, 21 L. Ed. 2d 637 (1969), which implements Aguilar v. Texas, 378 U.S. 108, 84 S. Ct. 1509, 12 L. Ed. 2d 723 (1964).2 Spinelli requires that two prongs be satisfied. U. S. v. McLeroy, 584 F.2d 746, 748 (CA5, 1978); U. S. v. Colon, 559 F.2d 1380, 1383 (CA5, 1977); U. S. v. Montgomery, 554 F.2d 754, 756-57 (CA5), Cert. denied, 434 U.S. 927, 98 S. Ct. 409, 54 L. Ed. 2d 285 (1977). The tip must contain information giving sufficient grounds to the officer relying thereon to conclude that the person giving the tip is credible or that his information is reliable. This is the "credibility" prong is the information believable?3 Second, the tip must contain a sufficient statement of the underlying circumstances from which the informer drew his conclusion that the suspect was engaged in criminal conduct. To avoid confusion arising from use of the word "reliable" in differing contexts, this might be called the "criminal conduct" prong, i. e., sufficient revealed circumstances to indicate that the informer could conclude that the suspect was engaged in criminal, as opposed to innocent, activity. The government acknowledges that the tips in this case do not satisfy the two-prong test. It relies upon on-the-scene corroboration of details in the letters.

The role of corroboration in boosting an inadequate tip over the probable cause threshold has followed a tortured and uncertain path. Under Spinelli it is permissible to add corroboration to the probable cause assay. But there is no probable cause unless the tip when combined with corroboration "is as trustworthy as a tip which would pass Aguilar 's tests without independent corroboration." Spinelli, supra, 393 U.S. at 415, 89 S. Ct. at 588, 21 L.Ed.2d at 643. The cases give us numerous instances of an inadequate tip being lifted over the threshold by corroboration of facts which of themselves are direct or circumstantial evidence that criminal activity is afoot. The suspect's own actions undergird the informer's conclusion that the suspect is engaged in criminal activity, the "criminal conduct" prong.4 For example, in U. S. v. Brand, 556 F.2d 1312 (CA5, 1977), Cert. denied, 434 U.S. 1063, 98 S. Ct. 1237, 55 L. Ed. 2d 763 (1978), the court recognized that an affidavit containing only the assertion that a named informer claimed that cocaine was located at a given address failed to pass the Aguilar-Spinelli test, but when police officers responding to a drug overdose emergency saw hypodermic needles, pills and marijuana on the premises where the cocaine was allegedly located, probable cause was established. In U. S. v. Brennan, 538 F.2d 711 (CA5, 1976), Cert. denied, 429 U.S. 1092, 97 S. Ct. 1104, 51 L. Ed. 2d 538 (1977), it was not until the police had observed suspicious activity that probable cause was established. "At the point at which the law enforcement officials detected the Beagle aircraft proceeding down the taxiway in the dark with its lights off, at a time almost exactly that predicted by Dufresne (the informer) in his estimate for the time required for a smuggling flight, the quantum-of-knowledge ring closed around Brennan." Id. at 721. In U. S. v. Tuley, 546 F.2d 1264 (CA5), Cert. denied, 424 U.S. 837, 98 S. Ct. 128, 54 L. Ed. 2d 99 (1977), probably the furthest this court has gone in allowing corroboration of innocent details to bolster an inadequate tip, the court found significant that police observed the defendants visiting the apartment of a known drug smuggler. Id. at 1268.

The matter becomes more difficult when the corroborated facts concern activity wholly innocent. The corroboration will tend to demonstrate that the informer is credible, that is, he accurately relates facts. But can verification of described or predicted activity which in itself is innocent verify that the informer had sufficient basis to conclude that the suspect was engaged in criminal activity? We expressed our doubts but fell short of a square holding in U. S. v. Brennan, supra at 720: "(a)lthough there is no consensus on the type of information required to corroborate an informant's tip, it is generally agreed that the better practice is to obtain corroboration of incriminating details." (footnote and citations omitted). Later, in U. S. v. Afanador, 567 F.2d 1325 (CA5, 1978), we found "innocent fact" corroboration sufficient to satisfy the reasonable suspicion necessary for a border search but noted that it was not sufficient to bring the informer's tip up to the Aguilar-Spinelli Test for probable cause (and that, indeed, the government conceded that it could not). Id. at 1328.5

A full understanding of "innocent facts" as corroboration begins with Draper v. U. S., 358 U.S. 307, 79 S. Ct. 329, 3 L. Ed. 2d 327 (1959), decided before Aguilar and Spinelli. In Draper police had corroborated all of the details of a previously reliable informer's tip, except for the allegation that Draper was carrying heroin. Draper is, we are told by the Spinelli Court, still good law. Examining Draper in Spinelli terms, we see that the credibility prong of Aguilar was satisfied by the arresting officer's testimony that he had received reliable information from the informer on various occasions. The criminal conduct prong was met by the corroboration of accumulated details, which gave the officer reasonable grounds to conclude that the crucial unverified fact in the tip that the suspect's conduct was criminal was sufficiently reliable. 358 U.S. at 313, 79 S.Ct. at 333, 3 L. Ed. 2d at 332.6 Thus, after Spinelli, Draper stands for the proposition that a tip from a Credible informer that does not reveal how he acquired his information may nevertheless satisfy the criminal conduct prong of Aguilar if the tip provides sufficient detail that one may reasonably infer, simply because the informer has so much correct and intimate information about the suspect's activities, that he is accurate in describing the activities as criminal. This is the "inside track" or "pipeline" theory, U. S. v. Tuley, supra at 1273 (Godbold, J., dissenting); U. S. v. Montgomery, supra at 757-58. It gives the assurance that the tip is not based on casual rumor, which is the danger that the reliability prong of Aguilar is intended to guard against. See Spinelli, supra, 393 U.S. at 416, 89 S. Ct. at 589, 21 L.Ed.2d at 644. Where there is such a wealth of detail concerning innocent activities, accurately given by a believable informer, Draper can lift the case over the threshold. There is enough to "permit the suspicions engendered by the informant's report to ripen into a judgment that a crime probably was being committed." Id. at 418, 89 S.Ct. at 590, 21 L. Ed. 2d at 645.

However, the fact that innocent details were accurately described may not be, by itself, enough to find a reasonable basis for concluding that the tipster has sufficient basis for the crucial fact that the suspect is engaged in criminal activity. In Draper the credibility prong was satisfied by the informer's previous track record as "reliable." Because there was strong enough basis to believe that the informer was a truthful relater of facts it was reasonable to make the leap from accurate description of innocent details to inferring that there was an underlying basis for the relater's conclusion that the suspect's activities were criminal. The two prongs of Aguilar are separate, and each serves a different purpose in assuring that probable cause is present, but they are interrelated and to a degree may be overlapping. A particularly strong showing on one prong may compensate for a weaker showing on the other. U. S. v. Sellars, 483 F.2d 37, 41 (CA5, 1973), Cert. denied, 417 U.S. 908, 94 S. Ct. 2604, 41 L. Ed. 2d 212 (1974).

With these general principles as our guide we turn to the corroboration of the tip in this case to determine whether it raises the inadequate tips to the level of compliance with Aguilar-Spinelli standards. Our approach is a functional one, in an effort to assure that the purposes of Aguilar are fulfilled while not creating technical requirements that impede law enforcement and serve no other purpose. The corroborated details include:

Identification of one of the driver-participants

Presence with Villarreal of his pregnant wife and child

Use of the red Nova and brown Mercury, each with license number as stated

Passage through the Sarita checkpoint travelling north, the direction of Houston

All of the facts corroborated by observation at the Sarita checkpoint, standing by themselves, are innocent in nature. Together, the corroborated facts paint an innocent picture of two cars heading north to Houston. The only source of suspicion directed at this compilation of innocent facts are the statements in the letters that the two cars would be smuggling heroin, buttressed by law enforcement experience that drugs are often smuggled through the Sarita checkpoint and that a common modus operandi for drug smuggling involves the use of several cars.

This case cannot get past the threshold. There is not such a wealth of accurate and intimate detail that we may infer a "pipeline." The given facts corroborated were of a public nature the makes of cars, their license plate numbers, the name of one of the drivers, and that the cars would be travelling through the Sarita checkpoint. The more intimate facts, such as packing contraband in the spare tires and exchanging them at roadside parks, were either not corroborated before the search or not corroborated at all. The informers were wrong on several of their details one of the drivers was identified as Montalvo; in fact the driver other than Villarreal was named Smith. Only two cars were used instead of the predicted three. The final letter said that the drivers passed through Sarita every few days; this tip was first received November 10 and defendants did not pass through until the night of November 19.

Had the author of either letter been shown to be a person who had given reliable information in the past, the scales might have tipped to probable cause.7 In U. S. v. Sellars, Supra, a warrant case, we said: "(i)n cases where the affidavit presents such cogent assertions of reliability the quantum of underlying circumstances which reveal the source of the informer's knowledge necessary to sustain the affidavit is clearly less than in cases where the indicia of informer reliability is less dramatic." Id. at 41; Accord, U. S. v. Anderson, 500 F.2d 1311, 1316 (CA5, 1974). The converse is equally true. Where the informer remains anonymous and as a result law enforcement officials have no idea whether he is credible there must be a strong showing of underlying circumstances that reveal the informer's source and significant corroboration of the tip. In this case the corroboration of innocent detail was not adequate to satisfy both the credibility and criminal conduct prongs.

REVERSED.

I cannot but agree with my Brother Godbold's excellent statement of the law for the majority. However, I dissent because I cannot agree with the majority's application of that law to these facts. I would hold that the Government Agents had probable cause to arrest the defendants and search their automobiles.

I agree with the majority that we must apply the two prongs of Aguilar1 and Spinelli2 to determine whether there was probable cause to make these warrantless searches. I agree with the majority that on-the-scene corroboration may buttress an otherwise deficient tip to justify a search on probable cause. I also agree with the majority that such corroboration may buttress both the "credibility prong" and the "criminal conduct prong."3 Finally, I agree with the majority that the "credibility prong" has been sufficiently buttressed here by corroboration. The majority and I part ways and take separate forks in the road in our conclusions about the "criminal conduct prong."

The majority defines the "criminal conduct prong" as "sufficient revealed circumstances to indicate that the informer could conclude that the suspect was engaged in criminal, as opposed to innocent, activity." Majority opinion at 938. The critical focus is on the reasonableness of the officer's conclusion as to the informant's state of mind, and this can be measured only by an objective appraisal of the facts upon which the officer has based such a conclusion. The court must determine if the officer could reasonably believe that the informant was sufficiently aware to conclude that crime was afoot. I disagree with the majority's opinion that "the corroborated facts paint an innocent picture of two cars heading north to Houston." Majority opinion at 940. Here the corroborated details include: (1) identification of one of the driver-participants; (2) presence with Villarreal of his pregnant wife and child; (3) use of the multiple-vehicle Modus operandi ; (4) identification of the make and models of the automobiles; (5) the license numbers of the automobiles; (6) passage through the Sarita checkpoint; (7) convoy directed toward Houston. In addition, the letters stated that crime was afoot an ongoing heroin smuggling operation. I believe that when the informant proved truthful about these corroborated details it was probable that the critical unverified facts were also true.4

I reach this conclusion in the entire context here. When corroboration consists of seemingly innocent conduct, more significance should be given to an informant's accurate prediction of future activity than would be given a conclusion of criminal conduct in the past.5 That is, in those instances in which the informant describes a scheme of events to take place at some point forward in time and conduct is subsequently observed which follows the scheme outlined by the informant, the reliability of the informant's assertion that the conduct is in aid of criminal activity is strengthened, even though the conduct itself in the absence of the tip would appear to "paint an innocent picture." In addition, the law enforcement experience in this area of our country is that drugs are commonly smuggled through these checkpoints in multiple vehicle convoys. This Court often has depended upon this fact of border life to help establish a reasonable suspicion to justify the stop of these convoys.6 Whatever argument may be made for the innocence of cars of a specific make, model and license number travelling in tandem through a border checkpoint, no one can seriously attribute to happenstance that the two described vehicles would be travelling at a specific interval, through a specified checkpoint on an unspecified day In the future. The majority is correct to point out that some details were not corroborated, either because the search took place before corroboration was possible, E. g., the planned rendezvous at the roadside park, or because the actual events differ from the prediction, E. g., the driver other than Villarreal was Smith and not Montalvo as named in the letter, although it was later determined that Montalvo was the registered owner of the vehicle. However, I do not think that the prediction necessarily was wrong regarding the timing of the heroin smuggling. The final letter which was received on November 10 stated that the convoy passed through Sarita every few days; the defendants were intercepted on November 19. This could mean that an intervening convoy simply went undetected. In any event, the corroborated accuracy of this predictive tip was not outweighed in my opinion. The reliability of the tip justified a conclusion that the informer was indeed aware of criminal activity afoot. I believe that the "criminal conduct prong" was satisfied. This being so, probable cause existed for the search and the exigency of the automobile exception to the warrant requirement justified these arrests and searches.

I dissent because I believe that the majority is guilty of the failing it professes to avoid: "creating technical requirements that impede law enforcement and serve no other purpose." Majority opinion at 940.