Hull v. United States — 404 U.S. 893

Dissent — Douglas J.

p. 893

No. 70-5331.

Hull v. United States.

dissenting.

On the morning of September 22, 1968, the petitioner and a companion, both Americans, were walking westward along Highway 94 near Tecate, California, and were three-fourths of a mile from the Mexican border. Two federal customs agents stopped them, learned their iden

p. 894

tities, directed the pair to walk in roadside dust, and concluded that their shoeprints were similar to those the officers had seen only hours earlier in sand “traps” some 50 yards south of the highway.^[1] Suspecting the two of smuggling contraband, the agents detained them while reinforcements arrived and backtracked in search of contraband which officers believed might have been' hidden along the pair’s path. During the probe two knapsacks containing marihuana were found at a spot 100 yards east of the point where the two had been stopped and near the trail which officers believed had been left by the petitioner and his companion. The officers then formally arrested them and afterwards traced what they suspected to be the pair’s footprints to a railroad tunnel nearer the border leading into Mexico.

The two were convicted of smuggling marihuana, in a trial before a jury. The Government was unable to produce direct evidence that the defendants had actually crossed the border via the railroad tunnel. In fact, all the agents could claim was that they had traced what they believed to be the defendants’ footprints into the mouth of the tunnel. Furthermore, no direct evidence was adduced to show, assuming the pair had crossed the border, that they had crossed with marihuana. Given the lack of direct evidence on any of these matters, however, the Government understandably desired to be more persuasive.

To shore up its case the Government introduced, over objection, testimony of an agent that the locality of these events had been “possibly the hottest spot on the Mexican border for smuggling.” He was permitted to tell of his participation in six prior, unrelated investiga

1 A “trap” is a sandy area brushed smooth at regular intervals and used to alert border patrols that persons have recently traversed the area on foot.

p. 895

tions of other persons’ smuggling activities. On four of these occasions, said the officer, arrestees had been tracked by their footprints and had been found possessing marihuana. On two other occasions, aliens had been caught unlawfully entering the United States. Apparently, the point the prosecution wished to convey was that using the tunnel was a favorite modus operandi for sneaking aliens and hemp into the United States. The Government now attempts to justify this testimony on the theory that it tended to rebut the possibility that the pair had merely been hitchhiking. As a second justification, the Government contends the remarks helped establish the “signcutting” credentials of the agent by showing that his tracking skills had proved successful on prior occasions.^[2]

1 would reverse the case and order a new trial.

The witness’ view on the criminal tendencies of persons found near the border was not relevant to any fact at issue in the trial. It was merely an attempt to construct a supposed class of suspect persons to which it appeared the petitioner belonged. Courts have been hesitant to admit statistical evidence of this nature because of the ease with which it can be abused.^[3] Only after many

2 “Signcutting” is the art of tracking the path of persons traveling in open spaces.

3 Judicial history is replete with examples of misapplied statistics in criminal eases. One prominent example was the conviction of Captain Alfred Dreyfus in France in 1899. He was accused of writing state secrets and conveying his inscriptions to German spies. The prosecution attempted to show that Dreyfus had penned the letter, by introducing lexicographical experts’ opinions. The prosecutor was successful in persuading the jury that there had been an “amazing” frequency of graphical similarities between the letters and others known to have been written by Dreyfus. But, as a panel of experts later pointed out, there was nothing statistically remarkable about the existence of the similarities. For the view that counsel for Dreyfus and the government commissioner failed *896 to comprehend any of the witness’ mathematical speculation, see A. Charpentier, The Dreyfus Case 52-53 (J. May transl. 1935). A more recent American case is that of People v. Collins, 68 Cal. 2d 319 , 438 P. 2d 33 (1968), in which the prosecutor at trial said he showed that only one in 12 million people could, as the defendant did, match the description of the perpetrator of a robbery. The California Supreme Court ultimately pointed out, among other criticisms of the mathematical demonstration, that in a population of 24 million the prosecutor’s assumption would lead to the conclusion that there were two persons matching the description, which alone should have established reasonable doubt. These cases and others misapplying statistics are discussed in Tribe, Trial By Mathematics: Precision and Ritual in the Legal Process, 84 Harv. L. Rev. 1329 , 1338 (1971).

p. 896

years of improvements were evidences of fingerprints, firearms ballistics, and radar data admitted in prosecutions. And, evidence of blood tests for identification purposes or polygraph readings still face an uncertain future in our courts.^[4] We have here an attempt by the Government to impute to a class of persons criminal tendencies simply on the strength of one custom agent’s anecdotal experiences in four prior investigations. The attempt was not relevant since a rational juror could not have sensibly used the information.^[5]

The jury might have suspected the two simply as marihuana users who had been hitchhiking along the highway and who had hidden the contraband upon seeing

4 As to the case of breathalizers, radar trackings of automobiles, polygraphs, blood tests for identification purposes, fingerprinting and ballistics tests, see C. McCormick, Evidence, c. 20 (1954); Pinkelstein & Fairley, A Comment on Trial By Mathematics, 84 Harv. L. Rev. 1801 (1971); Tribe, A Further Critique of Mathematical Proof, 84 Harv. L. Rev. 1810 (1971); Tribe, Trial By Mathematics: Precision and Ritual in the Legal Process, 84 Harv. L. Rev. 1329 (1971); and Finkelstein & Fairley, A Bayesian Approach to Identification Evidence, 83 Harv. L. Rev. 489 (1970).

5 See Committee on Rules of Practice and Procedure, Revised Draft of Proposed Rules of Evidence for the United States Courts and Magistrates, Rule 401, 51 F. R. D. 315 , 342.

p. 897

the customs agents’ automobile. Under these circumstances it would have been understandable for the Government to desire to show that a high proportion of persons found with hemp near the Mexican border in fact had recently brought their supply personally from Mexico. It may well be that such is the case. But that proposition does not follow from the four instances recalled by the customs agent. In Leary v. United States, 395 U. S. 6 , 30 (1969), we noted the increasing harvests of domestic crops of marihuana. And, of course, it would not follow that all possessors even of foreign grown cannabis personally import their supply. In short, the Government offered an unbalanced and incomplete picture of the ways in which persons under circumstances such as these had obtained their marihuana.

It is no answer that the defendant was free to challenge the Government’s general view of statistical probabilities by presenting other explanations of the circumstances or by impeaching the Government’s expert witness. We should not impose upon an accused the burden of independently generating probabilistic evidence and employing experts to study the criminal tendencies of a subgroup of the population. The Government should live up to higher standards and not be allowed to convict people of crimes on suspicion alone.

I would reverse and order a new trial.

Cases cited

Leary v. United States 395 U. S. 6 SCOTUS 1969-05-19 p. 897