This appeal is from a judgment of the United States District Court for the Northern District of Texas in favor of Appellee for personal injuries sustained while using a mixture of Appellants’ cosmetics. Appellee, Marjorie Ann Pruitt, sustained third-degree chemical burns on her scalp and right ear resulting from the application to her hair of a mixture of two products designed for bleaching purposes — “Helene Curtis New Blue Bleach,” manufactured by Appellant Helene Curtis, and “L’Oreal Creme Developer,” manufactured by Cosmair. The products were purchased from a beauty parlor in Terrell, Oklahoma, by a friend, Mrs. Hendren, who applied them to Mrs. Pruitt’s hair at Mrs. Hendren’s home in Terrell.
In response to special issues, the jury found that the Blue Bleach and Cosmair mixture “contained ingredients that were not suitable and reasonably fit for the purpose for which said products were used when used in combination with the other,” and that such ingredients were a proximate cause of the injury. Both products were found to have contained “corrosive substances” (defined as “any substance which in contact with living tissue will cause destruction of tissue with chemical action”) which were a proximate cause of the injury. Mrs. Pruitt was found to have followed the directions which accompanied the products and was found not to have been negligent in mixing or applying them. A verdict was returned in favor of Appellee in the amount of $64,500.00.
Appellants allege several grounds of error: (1) That the trial judge erred in not granting the motion for a directed verdict or motion for judgment notwithstanding the verdict since the evidence was insufficient as a matter of law to establish a defect in the mixture; (2) that the trial judge erred in not granting a directed verdict or judgment notwithstanding the verdict since as a matter of law Mrs. Pruitt was without the class of persons who could invoke the doctrine of strict liability against Appellants; (8) that the trial judge erred in submitting the special issues on corrosiveness; and (4) that the verdict was excessive. We agree with Appellants that as a matter of law the jury could not have rationally inferred that *848the mixture was defective for its intended use. Alternatively, we hold that Appellee was without the scope of the duty which the doctrine of strict liability has imposed on Appellants. Only these two points will be discussed.1
I. The Applicable Law
The mandate of Erie R.R. Co. v. Tompkins, 1938, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188, once again plunges this Court into the vexing and revolutionary field of products liability. The multi-state nature of the transaction necessitates a choice-of-law analysis. It is well settled that a federal court uses the substantive law of the state in which it sits and that under Klaxon Co. v. Stentor Electric Mfg. Co., 1941, 313 U.S. 487, 61 S.Ct. 1020, 85 L.Ed. 1477, substantive law includes the conflict-of-law rules. Although most courts in tort problems are moving away from the vested rights and territorial approach, Texas still follows the traditional lex loci delictus rule. Marmon v. Mustang Aviation, Inc., 416 S.W.2d 58 (Tex.Civ. App. — Austin 1967, no writ). Here no injustice results from the application of this rule because there is no real conflict. George v. Douglas Aircraft, 2d Cir. 1964, 332 F.2d 73, 76; Comment, 78 Harv.L.Rev. 1452 (1965). Both Texas and Oklahoma recently extended the scope of strict liability to encompass all products that are unreasonably dangerous. The Texas Supreme Court, relying on the Restatement of Torts § 402A, adopted the tort theory of strict liability. Shamrock Fuel & Oil Sales Company, Inc. v. Tunks, 416 S.W.2d 779 (Tex. 1967); McKisson v. Sales Affiliates, Inc., 416 S.W.2d 787 (Tex.1967), noted in 45 Texas L.Rev. 790 (1967). The basis for the Oklahoma decision, Marathon Battery Co. v. Kilpatrick, 418 P.2d 900 (Okl.1966), is not clear, but the decisions relied on justify the conclusion that Oklahoma will also adopt the theory of strict liability in tort.
The mere adoption of strict liability for all products, however, does little to facilitate a solution to the issues posed by this appeal. These issues are the nature and quantum of proof necessary to establish liability; a proper understanding of the concept of a defective product; and the scope of the maker’s duty to the consumer. Thus, as in Putnam v. Erie City Manufacturing Company, 5th Cir. 1964, 338 F.2d 911, 917, we must consider “all the available data,” including the restatements of law, treatises, law review commentary, and the majority rule.2 Hopefully, our “Erie prediction” will be indicative of what Oklahoma will do in the future and will not be easily erased. Ford Motor Company v. Mathis, 5th Cir. 1964, 322 F.2d 267, 3 A.L.R.3d 1002.
II. Policy Considerations
Initially, we review the policy considerations behind strict liability. With the technological revolution and modern marketing practices of this Century, Americans now enjoy the conveniences of many modern and beneficial products. These benefits to the many, however, have come at a high cost to a few. To combat the serious injuries
*849visited on this minority, the law has reexamined its traditional .reasons for imposing liability. This “rethinking” has caused many courts to abandon the traditional negligence analysis and impose liability without fault on the maker who puts the product into the stream of commerce.3 The justification for rejecting privity is based on the realization that our technological society, with its proliferation of products and mass advertising, demands judicial protection of the consumer who has neither the capacity nor opportunity to discover latent dangers in products. Henningsen v. Bloomfield Motors, Inc., 32 N.J. 358, 161 A.2d 69, 75 A.L.R.2d 1 (1960); Prosser, The Fall of the Citadel (Strict Liability to the Consumer), 50 Minn.L.Rev. 791 (1966). Several reasons have been advanced to justify this judicial pioneering: The maker impliedly represents that his product is fit for its intended use; consumer expectations are frustrated by harmful products; strict liability will make manufacturers more careful; and manufacturers are superior risk-bearers because they have the capacity to distribute the losses of the few to the many by the price mechanism. Wade, Strict Tort Liability of Manufacturers, 19 Sw.L.J. 5 (1965). Thus the risk of personal injury has become a cost of doing business. The more recent decisions have adopted a tort rather than a warranty approach, reasoning that recovery should not depend on the complex law of sales. E.g., Vandermark v. Ford Motor Co., 61 Cal.2d 256, 37 Cal. Rptr. 896, 391 P.2d 168 (1964); McKisson v. Sales Affiliates, Inc., 416 S.W. 2d 787 (Tex.1967). These decisions find the tort theory more appropriate because liability is imposed by law and because the maker should not be permitted to define the scope of its responsibility for defective products. Comment, Products Liability — Proceeding Apace, 33 Tenn.L.Rev. 341 (1966).
The particular market involved in this case is the cosmetics industry. In no other way can one glean a better conception of our “consumer perspective” than by considering the advertisements which constantly encourage women to beautify themselves. See 3 Frumer & Friedman, Products Liability § 2901 (1965); Cahn, Law in the Consumer Perspective, 112 U.Pa.L.Rev. 1 (1963). It is therefore understandable that courts have imposed strict liability on the cosmetics manufacturer who has bombarded the consumer with his impersonal merchandising techniques. Cowan, Some Policy Basis of Products Liability, 17 Stan.L.Rev. 1077 (1965). The mere imposition of strict liability on cosmetics makers does not, however, as the trial court’s charge assumed, mean that the maker is liable for any harm to anybody under any circumstances. In fact, the ambit of responsibility is more circumscribed. The maker is not an absolute insurer who is responsible for all physical hurts occurring in the course of using the product. Traynor, The Ways and Meaning of Defective Products and Strict Liability, 32 Tenn.L.Rev. 363 (1965); Freedman, “Defect” in the Product: The Necessary Basis for Products Liability, in Tort and in Warranty, 33 Tenn. L.Rev. 323 (1966). The question facing courts today is what doctrine will replace fault as a means of delimiting liability. The chief limitation (which was accepted by the Oklahoma Supreme Court in Marathon) is the requirement that the product be defective for its intended use. The Restatement of Torts § 402A provides a lucid definition: “One who sells any product in a defective condition unreasonably dangerous to the user or consumer * * * is subject to liabil*850ity for physical harm thereby caused to the ultimate user or consumer.”
Demanding that the defect render the product unreasonably dangerous reflects a realization that many products, such as cosmetics, have both utility and danger. James, Products Liability, Texas L.Rev. 114 (1955). Since, in the instant ease, there was no evidence of any miscarriage in the production and no foreign substance was found in either product, we are confronted with what has been termed a design problem: The product was exactly as intended and yet harm still occurred. See Wade, supra.4 For the design to be unreasonably dangerous, it must be so dangerous that a reasonable man would not sell the product if he knew the risks involved. Id. at 10. This definition demonstrates that the only change from the traditional negligence analysis is that the maker cannot be excusably ignorant of the defect; however, courts must still weigh the utility of the product against the risk of harm created. Id. at 16. The trial court’s submission of strict liability ignores these principles.
III. Federal Test for the Sufficiency of the Evidence
It should be evident from the foregoing definition of a defect that the test for the sufficiency of the evidence is of paramount importance. This Court recently reaffirmed in Planters Manufacturing Co. v. Protection Mutual Insurance Co., 5th Cir. 1967, 380 F. 2d 869, the principle that although state law governs the elements of a cause of action, the sufficiency of the evidence is a matter of federal law.5 This doctrine has been formulated to protect the seventh-amendment right to a jury trial by demanding uniformity in the exercise of the power of the trial judge to grant directed verdicts. 2B Barron & Holtzoff, Federal Practice and Procedure § 871.1 at 18 (Wright ed. 1961). Accordingly, since the federal policy favors jury trials, we must approach Appellants’ complaint that the trial judge erred in not granting these motions with the realization that they should be granted sparingly. Id. at 375.6
The test employed by the Fifth Circuit is that a fact issue must be submitted to the jury if reasonable men could differ on the conclusions to be reached from the evidence presented. Isaacs v. American Petrofina, 5th Cir., 1966, 368 F.2d 193; Wells v. Warren, 5th Cir. 1964, 328 F.2d 666. We must view the evidence and all reasonable inferences most favorably to the party against whom the motion is made. Moreover, only the evidence and the reasonable inferences which support Appellee’s theory may be considered. Wilkerson v. McCarthy, 1949, 336 U.S. 53, 69 S.Ct. 413, 93 L.Ed. 497; Berry v. United States, 1941, 312 U..S 450, 61 S.Ct. 637, 85 L.Ed. 945. Indeed after Planters our sole function is to ascertain if there is a rational basis in the record for the jury’s verdict. The Plant *851 ers decision, however, provides no explicit help in determining which inferences are rational. The polar ends of this problem may be stated as follows: Planters holds that an inference may be reasonable though based partly on conjecture;7 it is not, however, unconstitutional to direct a verdict for the defendant, Galloway v. United States, 1943, 319 U.S. 372, 63 S. Ct. 1077, 87 L.Ed. 1458, and it seems well settled that an inference is unreasonable if it is at war with uneontradicted or unimpeached facts. See Parker v. Wideman, 5th Cir., 1967, 380 F.2d 433; 2B Barron and Holtzoff, supra, at 390. The proper blending of these two polar positions can only be accomplished by understanding that the basic question is one of policy: How far is the court willing to let the jury speculate? Comment, Directed Verdicts and the Right to Trial by Jury in Federal Courts, 42 Texas L.Rev. 1053, 1063 (1964).
In deciding how much the jury can speculate it is well settled that an appellate court cannot weigh the evidence or the credibility of the witnesses. Wells v. Warren, supra. The line of demarcation which we are required to walk is ephemeral: We must conclude that an inference is unreasonable without falling into the trap of weighing all the evidence and deciding that while the jury’s inference is reasonable, the evidence shows that another inference is just as reasonable, if not more so. 2B Barron & Holtzoff, supra at 392. It is only to that limited extent that all the evidence is considered.
IV. Proof Required for Recovery Under Strict Liability
The doctrine of strict liability only removes the requirement of privity of contract; it does not prove Appellee’s case. Prosser, The Fall of the Citadel (Strict Liability to the Consumer), 50 Minn.L.Rev. 791, 840 (1966). Since there was no direct evidence of an identifiable defect, we must decide the meaning of a rational inference in a products liability case which rests entirely on circumstantial evidence. The breach of duty which the jury would have had to infer was that the New Blue Bleach and L’Oreal were unreasonably dangerous for their intended use. As stated, in order to justify that finding the jury must have concluded that the products were so dangerous that a prudent maker would never have marketed the products had he known of their condition. Such a conclusion would be reasonable only if the jury could have concluded that the mixture was defective ; that the products were intended to be mixed; that the application by Mrs. Hendren comported with professional standards; and, finally, that Appellee’s scalp was not hypersensitive. We have not considered the issue urged by Mrs. Pruitt that she was inadequately warned of danger. The jury found that the warnings to Appellee for use were inadequate, but that this failure was not the proximate cause of the injuries. Since Appellee failed to move for a directed verdict in the trial court, she cannot now challenge the sufficiency of the evidence. See 2B Barron & Holtzoff, supra at 424.
*852A. Appellee’s proof about the mixture.
First, we consider the proof offered by Appellee on the dangerous qualities of the mixture. Appellee elicited testimony that the directions accompanying the Helene Curtis bleach were followed.8 *853Most important is the fact that Mrs. Hendren checked Appellee’s scalp and found no cuts, eruptions, abrasions, or other abnormal conditions which rendered the hair unsuited for bleaching; that the hair had not been shampooed, brushed or massaged in the twenty-four hours prior to treatment; that a preliminary strand test showed no harmful effects to the hair by the mixture; and that the proper proportions of the products were used. Appellee also had evidence about the nature of the bleaching process. It was shown that the mixture of the two products produced a liberation of oxygen from the hydrogen peroxide. This process burned the hair pigment until it was colorless. It also had the potential for causing harm if the mixture was too toxic or left on the hair too long. Mr. Anderson, an analytical chemist, testified about the chemical qualities of the mixture. He had conducted experiments on the products to ascertain the active ingredients.9 He admitted that there was a chemical burn, but he was unable to say what caused the burn. He did testify that the mixture would have had to be much stronger than the ones he tested to cause Appellee’s burn. His conclusions were based on the fact that the concentration of the peroxide, which he thought was the dangerous ingredient, was normal. Finally, he stated that he did not think a bobby pin or cut on the head could have increased the activity of the products enough to cause this burn.
Appellee also obtained the statement from Mr. Dieter, a cosmetic chemist for Helene, that in order to produce such a burn the products either had to be defective or the directions not followed. Dr. Majors, Appellee’s personal physician, testified that in his opinion the burn happened suddenly. By this statement he meant about fifteen minutes. It was shown that the solution was on Mrs. Pruitt’s head about fifteen to twenty minutes before the burning started. Dr. Majors also testified that he did not believe a cut could produce this burn. His subsequent testimony is the most interesting. He stated that he thought one of two things happened: There was either some mislabeling of one of the bottles so that the solution or the powder was not what was usually in the products,' or the directions were not followed properly. He concluded that it was his opinion that this “chemical burn was produced by the application of the bleaching substance she told me she used on her scalp.”
On the basis of this testimony, Appellee asserts that the defect in the mixture speaks for iteslf. Although the’trend has certainly been to allow more and more circumstantial evidence to serve as the basis for liability, we believe that in this instance a finding of a defect was unwarranted. The evidence about the character of the mixture and its propensities for harm does not suggest a defect. The cornerstone rule in products liability is that proof of mere injury furnishes no rational basis for inferring that the product was defective for its intended use.10 Appellee has attempted to skirt that principle by invoking the doctrine of res ipsa loquitur. Certainly this doctrine has produced the most situations in which the jury has been allowed to infer negligence from an unexplained accident. Several elements must be present before the doctrine is applicable: The accident must be one which would not ordinarily occur without the presence of negligence; the plaintiff must show that the product was properly handled after it left the maker’s possession; and, finally, plaintiff must prove that the product has not been substantially altered after leaving the mak*854er’s control.11 Generally, the plaintiff is required to exclude these other causes by a preponderance of the evidence. Dement v. Olin-Mathieson Chemical Corp., 5th Cir. 1960, 272 F.2d 76; Ozark v. Wichita Manor, Inc., 5th Cir. 1958, 252 F.2d 671. This preponderance-of-the-evidenee test was employed because the courts looked to state law for the applicability of the doctrine. Since this state rule conflicts with the federal test, this Court will no longer rely solely on state law to determine the applicability of res ipsa loquitur. See 2B Barron & Holtzoff, supra, at 18. All that is in issue is the sufficiency of the evidence to infer a defect in a product.12 The factors considered in deciding whether the doctrine applies are still relevant; the only thing that has changed is the quantum of proof necessary to negate the other explanations of the injury13
We agree with Appellants’ assertion that when circumstantial evidence is the only proof, courts have infrequently inferred negligence (here a defect) simply from the accident and proof of careful conduct by the plaintiff, and then only in instances where the accident is the type which, standing alone, points an accusing finger at the maker. Furthermore, in all these eases the chance of mishandling is improbable.14 It is a matter of common knowledge that cosmetics can never be made completely safe for all users and can cause injuries for many reasons other than a defect in the product. Since injuries can be caused by the toxicity of the chemicals, contact with other products, the method of application, or an allergic condition, the circumstances do not point an accusing finger at the manufacturer. See Ravo v. Lido, 17 A.D.2d 476, 236 N.Y.S.2d 135 (1962).
In the instant case Appellee has failed to prove that it was the toxicity of the products which caused the injury rather than the other factors just enumerated. The testimony about the burning quality of the products is patently inadequate. This burning quality resulted from the oxidation process which occurred when the bleach and peroxide were mixed. No liability can be based on this burning potentiality since the very process of bleaching hair involves a chemical reaction or oxidation process by which the color in the hair is destroyed. It takes a powerful chemical to accomplish the decolorization. As one court observed, the sale as a bleaching fluid of a product that did not bleach would itself constitute a breach of warranty. Thus *855the very ingredient and quality which makes the fluid harmful if improperly used is the ingredient and quality which makes it fit for the purposes for which it was purchased. Landers v. Safeway Stores, Inc., 172 Or. 116, 139 P.2d 788 (1943). When products contain dangerous ingredients which are natural or inherent, courts have usually held that the product is not defective. Zorger v. Hillman’s, 287 Ill.App. 357, 4 N.E.2d 900 (1936). It is only when the manufacturer could not reasonably expect the consumer to discover the natural substance in the final product that liability has resulted. Bennett v. Pilot Products Co., 120 Utah 474, 235 P.2d 525, 26 A.L.R.2d 958 (1951); Betehia v. Cape Cod Corp., 10 Wis.2d 323, 103 N.W.2d 64 (1960). In the instant case we cannot say that Appellee did not know that the bleaching process and its natural ingredients were intended to decolor her hair.
The other proof on the character of the mixture is just as equivocal. See e. g., Rexall Drug Co. v. Nihill, 9th Cir. 1960, 276 F.2d 637; Benavides v. Stop & Shop, Inc., 346 Mass. 154, 190 N. E.2d 894 (1963). The two statements that the products were either defective or misapplied do nothing to exclude each other. Dr. Major’s statement that the “application of the bleaching solution” caused the burn is the most interesting. In Zampino v. Colgate-Palmolive Peet Co., 8 A.D.2d 304, 187 N.Y.S.2d 25 (1959), the plaintiff obtained a similar statement from a doctor to show that the aluminum sulfate in Veto deodorant was harmful. The court held that the statement amounted to no evidence of the fact that some ingredient in the product was harmful. Of course the application of the bleaching solution in the instant case caused the injury. The only circumstance in which this statement could conceivably support recovery would be one in which res ipsa loquitur was applicable. Here, however, it provides no basis for the inference that it was the nature of the mixture rather than the method of the applier which caused the burn. Hanrahan v. Walgreen Co., 243 N.C. 268, 90 S.E.2d 392 (1955).
Nor do all the statements stand alone. We present the following discussion as additional support for our holding that as a matter of law these products were not in a defective condition which was unreasonably dangerous for their intended use. Appellee’s problem is that she has not presented enough facts to constitute a cause of action in strict liability. Cf. Fidelity and Casualty Co. of New York v. Funel, 5th Cir. 1967, 383 F.2d 42. In all the other cases considered, the proof presented by the plaintiff was sufficient to make out a cause of action, despite conflicting evidence. Dick v. New York Life Ins. Co., 1959, 359 U.S. 437, 79 S.Ct. 921, 3 L.Ed. 2d 935; Berry v. United States, 1941, 312 U.S. 450, 61 S.Ct. 637, 85 L.Ed. 945. Our approach does not weigh the evidence or the credibility of the witnesses; we simply assert that the evidence about the characteristics of the mixture is not sufficient to constitute a cause of action in strict liability.
B. The Intended Use Doctrine: The Illegal Mixture.
In the first place, Mrs. Pruitt has not established a cause of action because she has not shown that the mixture of these products was intended. This fact had to be established because a product is not defective unless injury occurred during an intended use. Lartigue v. R. J. Reynolds Tobacco Co., 5th Cir. 1963, 317 F.2d 19; Spruille v. Boyle-Midway, Inc., 4th Cir. 1962, 308 F.2d 79. The directions on each package suggest that the mixture was not intended. The “New Blue Bleach” said: “Do not mix Helene Curtis New Blue Bleach with anything except Helene Curtis Creme Developer (or a good-grade or fresh 20 volume hydrogen peroxide).” It is undisputed that Cosmair peroxide is a cream rather than hydrogen peroxide. The label on the Cosmair bottle said: “For tinting or bleaching, use Oreor exactly as you would use 20-volume peroxide with Imedia Petite, Imedia Petite Excellence, Imedia Creme, Super Blue Creme Oil Light*856ener.” It is undisputed that Helene’s product is none of the above.
Mr. Dieter, the cosmetic chemist for Helene, compared the peroxides. He stated that hydrogen peroxide has the viscosity of water while the cream peroxide is similar to hand lotion. The cream peroxide contains additives which thicken the product and also provide some emollients or conditioning actions when used on the hair. He stated further that there was definitely a physical difference between the products, but his company had no way of ascertaining the chemical differences since the formula for the Cosmair product was a trade secret. Because of the uniqueness of cream peroxide, Mr. Dieter stated that Helene recommended that a hydrogen peroxide be used when Helene’s own Cream Developer was not used, since there is nothing unique about a hydrogen peroxide. Cosmair’s expert agreed that the trade secrets made it impossible to know about the products’ compatibility; that the directions were meant to be followed; and that he did not recommend that the products be used together. Appellee has suggested that these directions serve both commercial and safety ends. Although they may have a dual nature, the decisions demonstrate that a failure to follow directions cannot support a finding of a defect. Prosser, The Fall of the Citadel (Strict Liability to the Consumer), 50 Minn.L.Rev. 791, 824 (1966). Indeed, instructions accompanying cosmetics are an integral part of the warranty and are to be strictly followed. E. I. DuPont De Nemours & Co. v. Baridon, 8th Cir. 1934, 73 F.2d 26. This principle is particularly applicable to a bleaching product which has the inherent danger of burning the scalp or hair if misapplied in any way. Furthermore, directions become more important when products are to be mixed. All the relevant decisions presented evince a hesitancy to hold the maker responsible when mixtures are involved. No doubt many of the plaintiffs’ problems in these cases stem from the requirement that the other possible causes be negated by a pi’eponderance of the evidence. But aside from the preponderance-of-the-evidence approach, which we have rejected, these decisions reflect the valid policy concern for the potential abuses of jury speculation when mixtures are the cause of injury. What these decisions justifiably demand is that there be a rational basis in the record from which the jury can make a selective choice between the products before the manufacturer can be held responsible for the harm caused by his product or the mixture. See Bathory v. Procter & Gamble Distributing Co., 6th Cir. 1962, 306 F.2d 22; Sheptur v. Procter & Gamble Distributing Co., 6th Cir. 1958, 261 F.2d 221; Harrod et al. v. Edward E. Tower Co., 346 Mass. 532, 194 N.E.2d 392 (1963). This requirement is all the more warranted when the products involved were never meant to be used together. Unlike the case of Ozark v. Wichita Manor, Inc., 5th Cir. 1958, 252 F.2d 671, in which component parts were used to produce a final product, it is clear in the instant case that the products involved were not marketed for that purpose.
Nor can it be said that Appellants should have foreseen the mixture, since the trade secrets of one maker prevented any safety tests on the product of the other. Manufacturers cannot be responsible for a combination which they did not recommend and which they had no way of guarding against at the manufacturing stage. Lartigue v. R. J. Reynolds Tobacco Co., 5th Cir. 1963, 317 F.2d 19. Moreover, the mixing of these products amounted to an abnormal handling or substantial alteration which, because it was uintended and unforeseen, excuses the makers from responsibility for any harm. Noel, Manufacturers Negligence of Design or Directions for Use of a Product, 71 Yale L. J. 816 (1962); Prosser, The Fall of the Citadel (Strict Liability to the Consumer), 50 Minn.L.Rev. 791 (1966).
C. The Intended Use Doctrine: The Intended User.
Appellee has argued that she did not know the difference between hydrogen and cream peroxide. This *857argument reveals her second misconception about proof of a defect. Mrs. Pruitt has assumed that this product was made for her use and has argued that the directions were inadequate for the ordinary consumer. The uncontradicted evidence, as subsequently shown, confirms, however, that these products were intended to be used by a professional beautician. Since the decisions indicate that the maker is liable only to those whom it can reasonably expect to use its product, Mrs. Pruitt cannot recover. Merrill v. Beaute Vues Corp., 10th Cir. 1956, 235 F.2d 893; Bronson v. J. L. Hudson Co., 376 Mich. 98, 135 N.W.2d 388 (Mich.1965).
Appellee’s counter argument reveals the final fallacy in her position. Mrs. Pruitt asserts that a professional beautician in Terrell had mixed these two products and that the professional beautician who testified had done the same. Before these facts can become important, Appellee must show that Mrs. Hendren’s method of application met professional standards. Our requirement that Appellee prove that Mrs. Hendren applied the mixture as a professional is justified by the conclusion that these were not over-the-counter products, nor was Mrs. Pruitt the intended user. Accordingly, evidence which shows that harm was suffered when th: products were used by a layman does not necessarily demonstrate that' the products would have been defective if applied by a professional. Parker v. State, 201 Misc. 416, 105 N.Y.S.2d 735, 741 (1959). This proposition is valid since it is well settled that a product cannot be deemed defective if proper application would eliminate the possible dangers.
Mrs. Hendren had seen the “FOR PROFESSIONAL USE ONLY — NOT FOR PUBLIC SALE” language on the Helene bleach. Her understanding of this language was that if you did not know how to use these products you had better leave them alone. (R. 184) She admitted that she had no professional training. However, she had once lived near a beautician student and had used her books and practiced on her. She said she had used the mixture in question on two previous occasions. She related the following events concerning the accident: Before applying the solution she had checked the scalp and found no cuts or abrasions. She admitted that the solution touched the ears but contended there was no way to avoid the contact. Furthermore, she had tried to wipe off the solution whenever it touched the skin. The solution was first applied to the outer parts of the hair. At that time she tested the hair and decided it was not bleaching very fast, so she went to fix some iced tea. Mrs. Hendren finished her tea and then proceeded to apply the solution down to an inch of the scalp. It was during this final application that Mrs. Pruitt complained of a burning sensation. Mrs. Hendren immediately washed off the solution and called Dr. Majors.
On the basis of this testimony, the jury concluded that Appellee and Mrs. Hendren followed the directions contained on the products when mixing and applying them. The uneontradicted testimony of a professional beautician, however, was that a professional would have taken precautions in addition to those prescribed by directions. This evidence was elicited from Mrs. Nickerson, an instructor at Bud’s Hall Beauty College. This witness’ conception of a “professional application” is far different from the picture drawn by Mrs. Hendren of two friends reading the directions from only the Helene package and, drinking iced tea as the solution was applied. In order to acquire a license, a beautician must complete a course of one thousand hours, 475 of which are related to the process of bleaching hair. According to the professional witness, burning is the most common danger and one must therefore take precautions against leaving the solution on too long. A professional never depends on the customer’s complaints and constantly surveys the hair to detect burning. This burning can be detected by holding the hand over the head. Burning can also be seen when the solution starts to move away from the scalp or bubble. The professional would also have taken a strand test every five min*858utes to determine whether the hair and the solution were reacting favorably. On a virgin head, like Mrs. Pruitt’s, the solution should be removed and a new application used before the mixture is applied next to the scalp. In order to prevent the mixture from touching the ears, cotton could be packed around them. Finally, the theory book used by Mrs. Nickerson provided that a patch test should always be taken of a virgin head to determine whether the customer might be allergic to the product.
These additional safety precautions would have materially lessened the chance of injury. It is this uneontradicted testimony which nullifies the statements that the products were either defective or misapplied. Moreover, these additional safety precautions illustrate that written directions do not give all of the safety procedures which the intended user would have applied. Appellants were justified in anticipating these professional methods, since it is well settled that a manufacturer has the right to expect that his product will be used in the normal and customary fashion. McCready v. United Iron & Steel Co., 10th Cir. 1959, 272 F.2d 700; Bennett v. Pilot Products Co., 120 Utah 474, 235 P.2d 525 (1951). When these products were marketed, the makers could only foresee that they would be applied by a trained beautician. Therefore, the directions had to be adequate only for the professional’s use. Taylor v. Jacobson, 336 Mass. 709, 147 N.E.2d 770, 775, 76 A.L.R. 2d 1 (1958). It is also clear that the maker must give directions which will be adequate in light of the intended users’ training, Patterson v. George H. Weyer, Inc., 189 Kan. 301, 370 P.2d 116 (1962), and does not have to warn against dangers which are generally known. Jamieson v. Woodward & Lothrop, 1957, 101 Ü.S.App.D.C. 32, 247 F.2d 23; Parker v. State, 201 Misc. 416, 105 N.Y.S.2d 741 (1951). Thus, Mrs. Hendren’s methods amount to the conduct of an atypical user and render the finding of a defect in the products irrational. This abnormal handling could not have been foreseen, and it was a substantial alteration of the method of application.
Finally, the problem of allergic users and hypersensitive people is a recurring one in the area of strict liability for cosmetics. As one commentator has asserted, products which involve chemical reaction never have any tests which can guarantee absolute safety to all the consuming public. White-more, Allergies and Other Reactions Due to Drugs and Cosmetics, 19 Sw.L.J. 76 (1965). Oklahoma law indicates that an allergic plaintiff cannot invoke strict liability. Merrill v. Beaute Yues Corporation, 10th Cir. 1956, 235 F.2d 893; John A. Brown Co. v. Shelton, 391 P.2d 259 (Okl.1964). Appellee has attempted to show that her skin is normal. None of her evidence was connected to the prior use of these products. Some courts indulge in a presumption that the skin is normal. Jacquot v. Wm. Filene’s Sons Co., 337 Mass. 312, 149 N.E.2d 635 (1958). Although this presumption would usually be reasonable, it cannot apply here. As admitted by Dr. Majors, the only way of ascertaining whether Mrs. Pruitt was allergic to this particular mixture would be a test using these products. This is the function of the patch test which the professional would have performed. In such a situation, to presume that the, skin is normal would be contrary to common sense. This conclusion is reinforced by the additional factors that the products were mixed contrary to directions and applied by a novice. Therefore, it cannot be said that Appellee has established any rational basis for excluding the possibility of a hypersensitive reaction.
On the basis of the foregoing factors, we hold that Appellee has established no legitimate basis from which the jury could rationally infer that Mrs. Pruitt negated the other possible causes of injury. Bronson v. J. L. Hudson Co., 376 Mich. 98, 135 N.W.2d 388 (1965); John A. Brown Co. v. Shelton, 391 P.2d 259 (Okl.1964). A federal court does not function as a mere automaton in protecting seventh-amendment rights. *859Instead, its paramount function is to see that justice is done. Reuter v. Eastern Air Lines, 5th Cir. 1955, 226 F.2d 443. Pursuant to that policy, we feel justified in holding that the evidence produced by Appellee would allow jury speculation beyond the limits envisioned by Planters. See Fidelity & Casualty Co. of New York v. Funel, 5th Cir. 1967, 383 F.2d 42; McPherson v. Tamiami Trail Tours, Inc., 5th Cir. 1967, 383 F.2d 527; Huffstutler v. Hercules Powder Co., 5th Cir. 1962, 305 F.2d 292; Thomas v. Atlantic Coast Line Railroad Co., 5th Cir. 1955, 223 F.2d 1; McMeekin v. Gimble Brothers, Inc., W.D.Pa. 1963, 223 F.Supp. 896.
V. The Scope of the Duty Imposed by Strict Liability
The favored policy of jury trials in federal courts compels articulation of the other basis for our holding that Appellee should take nothing. Even if the jury could have properly found the mixture defective, there are additional obstacles to recovery. Specifically, the question presented is whether Appellants are responsible for the sale by the beauty shop to Appellee. What must therefore be decided is whether Mrs. Pruitt was a foreseeable user and whether the sale by the shop was foreseeable.
There is no “true rule” which will prevent the prostitution of the doctrine of strict liability into a form of absolute liability under which the maker is liable for all accidents resulting from unknown causes. Indeed, whether the ambit of responsibility will be the same as it is in negligence cases is still unsettled. Courts have attempted to answer this problem by again resorting to the intended use doctrine. This doctrine is composed of two factros: The marketing scheme of the maker and foreseeability of harm. Courts have defined foreseeability as the kinds of risks which the enterprise is likely to create, Lartigue v. R. J. Reynolds Tobacco Co., 5th Cir. 1963, 317 F.2d 19, or those risks which are inherent in the proper use of the products for the purposes for which they were intended. Spruille v. Boyle Midway, Inc., 4th Cir. 1962, 308 F.2d 79. It is these normal consequences which are a cost of doing business.
The dual concept of appropriate use and foreseeability of harm is traceable to Henningsen v. Bloomfield Motors, Inc., 32 N.J. 358, 161 A.2d 69, 75 A.L.R.2d 1 (1960). The court asserted that the interests of the person who could reasonably be contemplated to use the product demanded that society remove the doctrine of privity as a defense. This decision rested on the policy that it would be unjust to allow the maker to create a demand for the product by mass advertising and then evade liability with the shield of privity. Jacob E. Decker & Sons, Inc. v. Capps, 139 Tex. 609, 164 S.W.2d 828, 142 A.L.R. 1479 (1942).
The scope of responsibility envisioned by Henningsen is still in conflict. The battle has been waged around the rights of the innocent bystander who is neither a consumer nor user of the products. Some courts have focused on the advertising and marketing rationales of strict liability to deny the bystander recovery. These courts reason that since the bystander is not in the distribution chain he has no reason to expect protection from harm caused by the product. To them privity has been abandoned only in the marketing sense and there must still be some nexus between the maker and the injured party besides the injury. Comment, Products Liability — Proceeding Apace, 33 Tenn.L.Rev. 341 (1966). The essential nexus is that the maker by his advertisements must have attempted to induce the injured person to purchase the product. Prosser, The Fall of the Citadel (Strict Liability to the Consumer), 50 Minn.L.Rev. 791 (1966). In support of this position, it has also been argued that the warranty of fitness does not extend to the public in general. Hahn v. Ford Motor Co., 256 Iowa 27, 126 N.W.2d 350 (1964). However, in Piercefield v. Remington Arms Co., 375 Mich. 85, 133 N.W.2d 129 (1965), the implied warranty was extended to cover a bystander who was hurt when the gun *860his brother was shooting blew up.15 The court in Piercefield reasoned that foreseeability of harm was the test for the scope of liability. Moreover, since the warranty was implied by law, the doctrine of the best-risk-bearer also justified imposition of liability when the injury was foreseeable. Mitchell v. Miller, 26 Conn.Sup. 142, 214 A.2d 694 (1965). Apparently the rationale is that the maker can know that not all of his products are perfect and in light of this expectation can foresee that people within the vicinity of expected use will be hurt if the product is defective. Ford Motor Co. v. Mathis, 5th Cir. 1963, 322 F.2d 267, 274; Cudmore v. Richardson-Merell, Inc., 398 S.W.2d 640 (Tex.Civ.App.— Dallas, 1966, writ ref’d n.r.e.); Comment, Products Liability — Proceeding Apace, 33 Tenn.L.Rev. 341, 355 (1966).
The intended marketing scheme is one basis for deciding which users can be foreseen. In the instant case the uncontradicted testimony shows that these products were not marketed for use by the ordinary consumer. The bleach was sold through the Beauty Division of Helene Curtis which sells directly or through distributors only to beauty shops. The Cosmair expert also stated that the cream peroxide was sold solely t'o beauty shops. Mr. Thompson, the manager of the beauty shop supply house for the Terrell area, confirmed the validity of these company policies. He is the sole Helene Curtis distributor in the Terrell area and has only one competitor for the Cosmair products. The professional beautician was also uncontradicted in her testimony' tnat these products were not for public sale. It is this scheme which made the professional the foreseeable user and, as stated previously, it was only to this professional that directions about proper use had to be given. Hubbard-Hall Chemical Co. v. Silverman, 1st Cir. 1965, 340 F.2d 402, 404.
Appellee has confused the doctrines of adequate directions and adequate warnings. Directions tell how to use the product efficiently while warnings tell the dangers involved. Dillard & Hart, Products Liability: Directions for Use and the Duty to Warn, 41 Va. L.Rev. 145, 147 (1955). Here, the directions on both products, including the technical terms referring to the peroxides, were adequate for the professional. It is only in relation to the professional that the adequacy of directions must be judged since a manufacturer does not have to give directions for use to a consumer when he has not employed his advertising to induce this consumer to make a purchase. E. I. DuPont De Nemours & Co. v. Baridon, 8th Cir. 1934, 73 F.2d 26; Dillard & Hart, supra, at 161. Consequently, Appellants were not required to give directions which would allow Mrs. Pruitt to use the products safely. Indeed, they could confine their products to beauty shops, and incur liability only if a consumer were injured when the solutions were applied there. Cf. Stottlemire v. Cawood, D.C. 1963, 213 F.Supp. 897. Beyond that context they could not foresee any other use, unless it were determined that they could foresee the resale by the beauty shop. Even then the only additional duty would be to give adequate warnings that the product was not for use by a non-professional beautician.16 The decisions dealing with the foreseeability of harm to allergic users are analogous. It is clear that the maker does not have to give directions which allow the allergic person to use the product. All that is required is a warning that the product is harmful to persons having a particular allergy. Wright v. Carter Products, 2d Cir. 1957, 244 F.2d 53, 57; Crotty v. Shastenberg’s New Haven, Inc., 147 Conn. 460, 162 A.2d *861513 (1960); Casagrande v. F. W. Woolworth, Inc., 340 Mass. 552, 165 N.E.2d 109 (1960); Esborg v. Bailey Drug Co., 61 Wash.2d 347, 378 P.2d 298 (1963).
Two decisions provide further support for holding that Appellee was not a foreseeable user. In Harper v. Remington Arms Co., 156 Misc. 53, 280 N.Y.S. 862 (Sup.Ct. 1935) the manufacturer sold shells which were made solely for testing guns. The shells were sold to the plaintiff, who did not realize that the words “Proof Load” meant that the shells had an abnormal explosive force. The court found no liability, reasoning that the manufacturer was required to foresee only that class of persons who in the usual course of business would come within the orbit of danger. It could not foresee resale of the shells by the dealers to anyone except an arms manufacturer. In Kaspirowitz v. Schering Corp., 70 N.J. Super. 397, 175 A.2d 658 (1961), the plaintiff purchased from a druggist a dandruff control product which was intended to be sold only by prescription. Although the druggist assured the plaintiff that there was no danger, harm still occurred. The plaintiff contended that it was reasonably foreseeable that the product would be used on the consumer market despite the warning to the druggist. The court disagreed, holding that the warning did not run to the plaintiff since the manufacturer had not promoted his product through advertising to the public. See Stottlemire v. Cawood, D.C. 1963, 213 F.Supp. 897.
Harper and Kaspirowitz suggest that a warning to the beauty shop terminated the makers’ responsibility for the future sale of the product. While it is certain that appropriate words of caution are a part of the obligations incurred in producing and selling commodities, Wright v. Carter Products, 2d Cir. 1957, 244 F.2d 53, the warning need only be reasonably calculated to reach and be understood by the person likely to use the product. James, Products Liability, 34 Texas L.Rev. 45 (1955). These words of caution have to be placed so as to catch the attention of the user and give a fair indication of the danger. Spruill v. Boyle-Midway, Inc., 4th Cir. 1962, 308 F.2d 79. The only beautician who testified knew that the products were not for public sale. The double warning on both sides of the Helene Curtis container imports this fact beyond any dispute: “FOR PROFESSIONAL USE ONLY — NOT FOR PUBLIC SALE.” This was the only warning read by the parties before use. Both Mrs. Hendren and Dr. Majors testified that they understood the labels meant that the products were not for one unskilled in beauty applications. The Cosmair label contained the warning on the front: “FOR PROFESSIONAL USE.” These words were sufficient, in light of the beautician’s training and status as an expert, to apprise any beautician that the products were not for amateurs and thus not for resale.17 Therefore, the resale by the beauty shop amounted to an intervening cause which severed the causal chain and insulated Appellants from liability for any subsequent harm. Mrs. Hendren’s failure to take the safety precautions an expert would have taken constitutes another intervening cause. The Restatement of Torts § 402A, which confines liability to products that have not been substantially altered after leaving *862the maker’s control, supports this holding. Cf. Grey v. Hayes-Semmons Chemical Co., 5th Cir. 1962, 310 F.2d 291; Mazzi v. Greenlee Tool Co., 2d Cir., 1963, 320 F.2d 821; Comment, Products Liability — Proceeding Apace, 33 Tenn.L. Rev. 341, 355 (1966). Surely this doctrine applies both to the product’s composition and to its method of distribution. The question of intervening cause has arisen often when the distribution chain is composed of the maker, several intermediate parties, and finally the consumer. These eases uniformly hold that when one of the intermediate parties is notified of the danger and proceeds to sell the product anyway, the manufacturer is insulated from liability.18
Although the foregoing considerations are sufficient to deny Appellee recovery, there are even more compelling reasons for our holding. These additional considerations grow out of the doctrine of foreseeability of harm. That doctrine is a judicial explanation of the social policies which limit liability. Pease v. Sinclair Refining Co., 2d Cir. 1939, 104 F.2d 183, 123 A.L.R. 933. Once fault is removed as a limitation on liability, these policy considerations should be candidly explained. Until Americans have a comprehensive scheme of social insurance, courts must resolve by a balancing process the head-on collision between the need for adequate recovery and viable enterprises. Wilson, Products Liability, 43 Calif.L.Rev. 809 (1955). This balancing task should be approached with a realization that the basic consideration involves a determination of the most just allocation of the risk of loss between the members of the marketing chain.
It has been argued that the manufacturer should absorb the “typical risks” of marketing his product. Wilson, Products Liability, 43 Calif.L.Rev. 809 (1955). Is the chance that a non-professional would use the product a typical risk? Several factors are pertinent: The nature of the ultimate transaction; the methods employed in distributing the product; and the type of activity in which the user was engaged when harm occurred. Seely v. White Motor Co., 63 Cal.2d 9, 45 Cal.Rptr. 17, 403 P.2d 145 (1965). In the instant case the products were marketed through distributors to innumerable beauty shops all over the country. The distributor in the Terrell area was clear that the products were to be sold only to beauty shops. The warnings on the packages supported this testimony. Indeed, the advertisements used by Helene to induce the purchase by the beauty shops stated: “YOU’LL SEE THE DIFFERENCE * * * YOUR PATRONS WILL FEEL THE DIFFERENCE WITH HELENE CURTIS NEW BLUE BLEACH.” In this type of market, it would be impossible to police the prohibited reselling by trade agreements. Nor can the single sale by a beauty shop in Terrell, Oklahoma be indicative of such a widespread practice in the industry that we can hold that the makers knew or should have known of the practice and should have taken additional precautionary steps. Noel, Manufacturers Negligence of Design or Directions for Use of a Product, 71 Yale L.J. 816 (1962). With only one injury since 1960, it would appear that reliance on the distributor to sell exclusively to beauty shops, the training of the beauticians, and the warnings on the package is adequate to prevent the mishandling of the products. James, Products Liability, 34 Texas L.Rev. 45, 48 (1955).
The type of activity in which the user was engaged when harm occurred also cuts against the makers’ re*863sponsibility. There is no doubt that the primary “assault on the citadel of privity” has been led by consumer pressure. The decisions have suggested that the maker should be responsible for harm only when a consumer who has been induced through advertising to purchase the commodity is injured. See Henningsen and Decker, supra. That factor is not present here. Indeed, it is clear that these products were not household goods and not intended for public sale. Furthermore, if the judicial opinions are in conflict on whether the public will be willing to absorb the cost of injuries to innocent bystanders, there can be no doubt that the public will be unwilling to pay higher prices for products when the injured plaintiff is an unforeseen and unauthorized user. Mitchell v. Miller, 26 Conn. Sup. 142, 214 A.2d 694 (1965). This consumer’s expectations would never have been frustrated if she had exercised any precaution in seeing whether this product was meant for her. Cf. Marathon, supra, 418 P.2d at 915. It is not unreasonable to demand that the user exercise minimum precaution. Betehia v. Cape Cod Corp., 10 Wis.2d 323, 103 N.W. 64.19 All prior cases have assumed that the consumer is incapable of finding out about the dangers. That rationale does not hold true here. See Sawyer v. Pine Oil Sales Co., 5th Cir., 1946, 155 F.2d 855. In fact, Appellee simply took a cursory glance at the directions on the Helene product and did not even read the Cosmair label. When the product is a bleaching solution specifically designed for burning all the hair pigment, surely the manufacturer can expect more self-protection than that. If Appellee chose to entrust herself to the care of a friend, she cannot hold the maker liable for the friend’s inadequacies. Therefore, although consumers’ expectations are relevant, the expectations frustrated in this case are not the ones which have been the impetus behind the imposition of strict liability. Alternative ways existed for Mrs. Pruitt to satisfy her desire to be a blonde. Cf. Comment, Strict Products Liability and’ The Bystander, 64 Colum. L.Rev. 916 (1964). She could have simply gone to a qualified beautician.
The paramount reasons for denial of recovery in this case are the principles which have been traditionally employed to invoke the doctrine of strict liability. These policies compel pur holding that the imposition of strict liability on these Appellants simply because they are the best-risk-bearers would be an undue impediment to business and an intolerable injustice. The policy that the maker has the greatest capacity to distribute the risks of loss is not such a blanket rule that it applies even though justice demands that another member of the marketing chain take the risk of loss for the injury. Confining the makers’ responsibility to harm incurred by use of the products in a beauty shop is not a revival of the doctrine of privity of contract. It is simply an attempt to confine the scope of liability to the zone of danger which could reasonably have been foreseen before these products were sold. Even the decisions which give the broadest definitions of the intended-use doctrine do not go as far as Appellee urges. The decisions have expanded the doctrine of intended use to include the environment in which the product is used and the incidental and attendant consequences that accompany normal use. Spruille v. Boyle-Midway, Inc., 4th Cir. 1962, 308 F.2d 79; Boyl v. California Chemical Co., D.Or.1963, 221 F.Supp. 669; Hardman v. Helene Curtis Industries, Inc., 48 Ill.App. 2d 42, 198 N.E.2d 681, 12 A.L.R.3d 1033 (1964); Haberley v. Reardon Co., 319 S.W.2d 859 (Mo.1958). The environment of the products in this case is a beauty shop. Their attendant use and incidental consequences do not include resale by the beauty shop, a mixture contrary to directions, and application by a *864novice. The risk created by this resale was neither the kind which the product was likely to create, nor inherent in proper use. Any other holding would twist the doctrine of foreseeability of harm into that degree of clairvoyance which hindsight alone can give. Butler v. L. Sonneborn Sons, Inc., 2d Cir. 1961, 296 F.2d 623; James, Products Liability, 34 Texas L.Rev. 45 (1955). Moreover, were we to bridge the gap between the present decisions and the instant situation, we could no longer say that manufacturers are not absolute insurers for all physical harm which occurs during use of the product. See Goldberg v. Killsman Instrument Corp., 12 N.Y.2d 432, 442, 240 N.Y.S.2d 592, 191 N.E.2d 81, 86-87 (1963); Calabresi, Some Thoughts on Risk Distribution and the Law of Torts, 70 Yale L.J. 499 (1961). The judicial doctrine of strict liability was not created to sanction such results. We therefore reverse the judgment of the district court and render judgment in favor of the Appellants.