McCORMICK, Justice.

Plaintiff Jon Kevan Chown appeals from adverse judgment in this products liability action which was tried to the court at law. The action was predicated on theories of design negligence and strict liability. In holding for defendant, the trial court found plaintiff failed to prove the elements of unreasonable danger and proximate cause. Plaintiff contends the trial court erred in both findings. Because we find no merit in plaintiff's attack on the unreasonable danger ground, we affirm without reaching the proximate cause ground.

This case arose from an industrial accident which occurred on July 21, 1975. Plaintiff had been employed for two weeks as a utility man at the Bandag, Inc., plant in Muscatine. He was being trained to work on a milling machine and a calender. On the occasion involved, he was working on the calender for the first time. A calender is a "machine with rolls used for frictioning, sheeting, coating, and spreading of rubber or rubber compounds." The rolls were like the rollers in a wringer washing machine. The calender in this case was manufactured by the Farrel Company, a division of defendant USM Corporation, sometime between 1900 and 1904.

The function of the calender was to compress a heated slab of rubber to a desired thickness in a continuous process. For that purpose, the rolls were heated to 245°. As the slab of rubber entered the machine on one side, the operator inserted polyethelene sheeting between the middle and bottom rolls on the other side to separate the compressed sheets of rubber emerging from the machine.

The accident happened as plaintiff was inserting the sheeting between the rolls. As he was attempting to straighten the sheeting with his left hand, his right glove and hand were caught between the rolls and pulled into them at the "nip point," the point at which the rolls came together. It took approximately 45 minutes to extricate plaintiff's hand. He suffered crush and burn injuries which required the hand to be amputated.

No guarding device existed to prevent the operator's hand from coming into contact with the rolls. Plaintiff contended the absence of a barrier guard to prevent the operator's fingers from being drawn into the rolls at the nip point constituted design negligence and a design defect for purposes of strict liability. In this appeal he alleges the trial court erred as a matter of law in each ground of its ruling.

Our review is guided by familiar principles. The trial court's findings of fact in a law action tried to the court have the effect of a special verdict of a jury. On review, the evidence is viewed in its light most favorable to the judgment. When the court denies recovery because of a party's failure to carry the burden on an issue, we will not interfere unless we find the burden was carried as a matter of law. Eldridge v. Herman, 291 N.W.2d 319, 321 (Iowa 1980). We are not bound, of course, by trial court determinations of law. Farmers Insurance Group v. Merryweather, 214 N.W.2d 184, 187 (Iowa 1974).

Two of plaintiff's assertions of error in the court's ruling on the unreasonable danger element relate to the weight of evidence and one concerns the effect of industry custom and practice on proof of the element.

I. Weight of evidence. The same argument concerning the strength of the evidence is involved under both theories of recovery. In order to prove design negligence, it was necessary for plaintiff to show the calender was unreasonably dangerous because of defendant's failure to use reasonable care in its design. See Bengford v. Carlem Corp., 156 N.W.2d 855, 864 (Iowa 1968); Restatement (Second) of Torts § 397 (1965). To establish strict liability, it was necessary for plaintiff to show the design of the calender was unreasonably dangerous. See Eickelberg v. Deere Co., 276 N.W.2d 442, 444 (Iowa 1979); Aller v. Rodgers Machinery Mfg. Co., Inc., 268 N.W.2d 830, 834 (Iowa 1978). Thus proof of unreasonable danger was an essential element under both theories.

The difference between the theories has been characterized as follows:

The essential difference between an action in negligence and one in strict liability (or breach of warranty) lies not in the condition of the product but in the requirement in the negligence action of additional proof regarding the nature of the defendant's conduct. In the negligence action, not only must the product itself be found actionable, but the defendant must also be found negligent in letting the product get into that dangerous condition, or in failing to discover the condition and take reasonable action to eliminate it. In strict liability, this is not required; all that the plaintiff must do is show that the product was in the dangerous condition when it left the defendant's control.

J. Wade, On Product "Design Defects" and Their Actionability, 33 Vand.L.Rev. 551, 553 (1980). That difference is not material here.

Because unreasonable danger was a common element, the trial court ruling on the failure of proof cannot be upset under either theory unless the evidence was strong enough to compel a finding of unreasonable danger as a matter of law.

The unreasonable danger element is explained in Aller, 268 N.W.2d at 834-36. One test of unreasonableness is whether the danger is greater than an ordinary consumer with knowledge of the product's characteristics would expect it to be. Another test is whether the danger outweighs the utility of the product. In a design case, the risk-utility analysis involves balancing of "the gravity of the danger posed by the challenged design, the likelihood that such danger would occur, the mechanical feasibility of a safer alternative design, the financial cost of an improved design, and the adverse consequences to the product and to the consumer that would result from an alternative design." Barker v. Lull Engineering Co., 20 Cal.3d 413, 431, 143 Cal.Rptr. 225, 237, 573 P.2d 443, 455 (1978). See Back v. Wickes Corp., ___ Mass. ___, 378 N.E.2d 964, 970 (1978).

Another way of framing the issue is to say the plaintiff must prove the design is "not reasonably safe." See Wade, supra, at 571. The element must be established as of the time the product was manufactured. Ward v. Hobart Manufacturing Co., 450 F.2d 1176, 1182 (5th Cir. 1971); Fincher v. Ford Motor Co., 399 F. Supp. 106, 114 (S.D. Miss. 1975); aff'd mem., 535 F.2d 657 (5th Cir. 1976); Aller, 268 N.W.2d at 837.

Plaintiff relied on the testimony of Donald E. Wandling, a consulting engineer who studied the calender after the accident. Mr. Wandling said the accident was caused by the absence of a barrier guard or other safety device which would protect the nip point of the rolls. He said a barrier could be constructed which would admit a sheet of material but would keep fingers away from the rolls. He said defendant's design drawings showed the barrier could have been put on the calender. Based on his research among articles in the American Machinist Magazine during the period 1910 to 1913, he said such devices were known in 1900-1904, when the calender was manufactured. He did not say they were used on calenders then but that the technology for their use existed.

Other evidence showed that barrier guards were not included in the design of defendant's calenders until 1923. The first safety standards requiring such guards were promulgated in 1927 by the American Standards Association under the auspices of the United States Department of Labor. Evidence was also received that the magazine articles relied on by Mr. Wandling concerned guarding devices on punch presses and machine tools rather than calenders. Even on those machines, according to a defense witness, the articles did not discuss barrier guards at nip points but described devices to protect the operator's hand in the machine itself.

Under this record, we cannot say the trial court was compelled as a matter of law to find the calender was unreasonably dangerous. The court was not required to find that the defendant in 1900-1904 could reasonably have foreseen that a consumer in 1975 would expect barrier guards to be included in the design. This is particularly true since the process in which the calender was being used by Bandag was not invented until approximately 1946. Furthermore, employing the risk-utility analysis, the court was not compelled to find that the safety device was technologically and practically feasible at the time of manufacture. The court could find, as it did, that the design of the calender was not unreasonably dangerous under 1900-1904 standards.

We have no basis for holding that plaintiff proved his design negligence and strict liability theories as a matter of law.

II. The effect of industry custom. A more difficult issue is presented by plaintiff's contention that the court reached its conclusion through application of an erroneous principle of law. This contention is based on the court's equating of the custom and practice of the industry with the state of the art in determining whether the design was unreasonably dangerous. The court said:

Notwithstanding the testimony of Mr. Wandling, which was based upon studies of old trade magazines and treatises, there does not appear to be anything in the record that would indicate that the calender as originally built was not consistent with the custom and practice of the industry during the period of 1900-1904. This would also mean that the manufacturing was consistent with the state of the art at that time.

Plaintiff points out, correctly we believe, that a distinction exists between custom of the industry and state of the art. Custom refers to what was being done in the industry; state of the art refers to what feasibly could have been done. The distinction is made in Hancock v. Paccar, Inc., 204 Neb. 468, 479, 283 N.W.2d 25, 35 (1979), as follows:

While the jury may consider, as evidence of the state of the art, the fact that no manufacturer is doing that which it is claimed could be done, such evidence will not establish conclusively the state of the art. Obviously, the inaction of all the manufacturers in an area should not be the standard by which the state of the art should be determined. Whether the design represents the state of the art is still a question of fact to be determined by the jury.

Judge Learned Hand in the case of The T.J. Hooper, 60 F.2d 737 (2d Cir., 1932), said: "Indeed in most cases reasonable prudence is common prudence; but strictly it is never its measure; a whole calling may have unduly lagged in the adoption of new and available devices. It never may set its own tests, however persuasive be its usages. Courts must in the end say what is required; there are precautions so imperative that even their universal disregard will not excuse their omission."

. . . . .

The question therefore is not whether anyone else was doing more, although that may be considered, but whether the evidence disclosed that anything more could reasonably and economically be done.

In the Uniform Product Liability Law developed by the Department of Commerce, the state of the art concept is embraced in the concept of technological and practical feasibility. See Uniform Products Liability Law, § 104(B)(2)(b), 44 Fed.Reg. 62714, 62721 (1979). State of the art usually means what is feasible; custom in the industry may be something less.

Not every safety device for which technology exists is necessarily feasible. See Hagans v. Oliver Machinery Co., 576 F.2d 97, 99-101 (5th Cir. 1978). However, we assume, without deciding, that the trier of fact could have found the barrier guard to be feasible at the time of manufacture in this case. Plaintiff's problem is that he did not urge a distinction between state of the art and industry custom at trial. The only discussion in the record of those concepts appears in deposition testimony of plaintiff's witness Wandling which was introduced by defendant. That testimony includes the following:

Q. Do you have any information with regard to the custom in the industry prior to 1927 with regard to furnishing devices of which you just made mention? A. No, I have not made a study of that.

Q. Do you have any information with regard to — and maybe it's the same question — the state of the art with regard to these safety devices back in — prior to 1927? A. No.

Q. Do you understand what I mean by "state of the art?" Is that a term used in the engineering field, or not? A. It is a term used in the engineering field, yes.

Q. What do you understand that to mean? A. Well, "state of the art" would be the practice of people involved in the manufacture of various kinds of equipment.

Thus plaintiff's witness equated state of the art with custom in the industry. Even if the record would have supported finding the barrier guard was a technologically and practically feasible design feature in 1900-1904, although not customary, the trial court obviously employed the term state of the art in the same way as plaintiff's witness. The court merely found that plaintiff did not prove the design was inconsistent with "the practice of people involved in the manufacture" of such machines. We do not believe the court can be put in error for using the definition of state of the art supplied by plaintiff's witness.

Furthermore, the trial court did not say plaintiff's failure to prove that the design of the calender was inconsistent with the custom and practice of the industry was conclusive on the issue of unreasonable danger. Rather, the court's finding merely shows the court recognized the relevancy of such evidence. Plaintiff acknowledges its relevancy, citing negligent design cases like Varas v. Barco Manufacturing Co., 205 Cal.App.2d 246, 22 Cal.Rptr. 737 (1962). See Restatement (Second) of Torts § 295A (1965). As demonstrated in Hancock, evidence of custom is similarly relevant on the issue of unreasonable danger, but not conclusive. 204 Neb. at 479, 283 N.W.2d at 35. We are unable to agree with plaintiff's contention that the trial court gave conclusive effect to the absence of evidence of deviation from custom.

Therefore, we find no merit in plaintiff's attack on the trial court's ruling finding a failure of proof on the issue of unreasonable danger. Because this ground independently supports the court's holding, we do not reach plaintiff's contention that the court erred in its holding on the issue of causation.

AFFIRMED.