This is an appeal from a jury verdict in favor of all defendants and against all plaintiffs, in an action for damages due to an explosion which occurred on February 12, 1953, at the Hercules Powder Company’s explosive manufacturing plant near Pinole, California. The appellants are Hercules Powder Company, joined pursuant to Labor Code, section 3853, by 34 individual appellants who are 22 injured Hercules employees and the heirs of 12 deceased Hercules employees. All of the appellants alleged negligence in the design, manufacture, installation, inspection and repair of an automatic sprinkler system, against the respondent, Automatic Sprinkler Corporation and two of its employees, J. W. Mott and F. E. Zaiser. Hercules alone also alleged breach of warranty. The lower court granted a non-suit to Zaiser as to all causes of action, and a nonsuit to Mott on the causes of action for breach of warranty. Respondents asserted the defense of contributory negligence against appellant Hercules and the defense of assumption of risk against the individual appellants. The facts are as follows:

In 1951 Hercules built a new “Pulp Dry & Dope House” as the previous one had been destroyed by fire in June 1950. In this building certain materials known as “dope” were processed. In another part of the Hercules plant, known as the “mix house” these dopes were combined with nitroglycerin to make various kinds of commercial dynamites.

In June 1951, after some correspondence and in accordance with Hercules specifications, the respondents installed a new type of “low pressure dry pipe” valve and automatic sprinkler system in the new dope house. The system worked as follows: Inside the enclosed valve, water under pressure at 90 pounds per square inch was held down by a cover or clapper; throughout the sprinkler piping and behind the sprinkler heads, air was held at a pressure of 6 to 7 pounds per square inch. When the air pressure above the clapper was reduced to 3 pounds per square inch or less, a weight mechanism in the valve released a latch which opened up the clapper, and permitted the water to pour through the vertical riser pipe at the top of the valve to the sprinkler piping and sprinkler heads onto the fire. The air pressure above the clapper was reduced by the melting of a fusible link in the sprinkler head. The fusion point of the link was 165° F. which is several hundred degrees below the burning temperature of the dopes processed. The design of the valve rendered the system inoperative if the valve was waterlogged by water leaking into the dry piping above the clapper and exerting water pressure equal to or above 3 pounds per square inch. At some time before the accident, Hercules installed a drip leg on the air pipe furnishing the air supply to the valve, to drain off any water which accumulated in the air pipe. One hundred and forty-seven sprinkler heads spaced 4 to 5 feet apart were installed near the ceiling of the building.

The respondent Automatic Sprinkler Corporation was informed that the sprinkler system and valve were to be installed in the dope house which was a place of “extra-hazardous occupancy,” but it does not appear that the respondent was consulted as to what type of sprinkler system and valve would be best for the particular building. On other occasions, Automatic had been consulted by Hercules about high speed fire prevention and extinguishing systems, and had installed many different types of devices in various Hercules plants. Hercules had been informed of the development of the dry pipe valve for a number of years and had been informed by Automatic that the system was reliable and would work within 15 to 20 seconds after the fusion of the sprinkler heads. The system installed by respondent was the first of its kind in California, and the only system of its kind on the market at that time. Respondent Mott, the district superintendent of Automatic Sprinkler Corporation, who was charged with the installation of the system had never seen or been instructed about the “low pressure dry pipe valve” which was to set the sprinkler system in motion.

Although the respondent had no contract for the maintenance of the sprinkler system with Hercules, respondent regularly inspected and tested the dry pipe valve, at intervals of about six months, after the initial installation which was completed in June 1951. There is conflicting evidence as to the amount and type of instruction given to Hercules employees by the respondents in the course of the installation and subsequent inspection of the system, or about the possibility of the water-logging of the valve.

In January 1953, a month before the accident, Mott again tested the valve in the presence of several Hercules employees. The apparatus failed due to the presence of several gallons of water that had collected above the clapper of the valve. In accordance with instructions he had received from the home office of Automatic, Mott installed a new latch in the clapper of the valve. There was some difficulty in the installation of the new latch. In another test the system worked in seven seconds. Mott then reset the system and checked to see that it was free of water. This was the last contact anyone had with the valve prior to the accident. Various Hercules employees, however, drained the drip leg pipe and checked the gauges of the valve. One Hercules employee was charged with cheeking the system. In several tests of the system conducted by Hercules, it worked normally.

In September 1952, there was a minor fire in the dope house caused by a welding spark during repair work. At this time the automatic sprinkler system did not function; several sprinkler heads were manually broken by Hercules employees, and worked but were not near enough to the fire. Thereafter, unknown to the respondent, Hercules had installed a “wet pipe” sprinkler system in the screen used to remove impurities from the “dope” because this screen was covered and could not be reached by the dry pipe sprinkler system. This sprinkler system apparently operated on February 12, 1953, as there was little evidence of fire in the screen. After the screening operation, the dope fell into canvas covered boxes on wheels measuring 34% by 37 by 26 inches known as “hods.”

Three types of dope were in the dope house on February 12, 1953: Yibrogel, which consists of 74.75 per cent sodium nitrate, 15.15 per cent wood pulp, 2.03 per cent starch, 5.05 per cent corn cob meal and 2.02 per cent ground limestone; three hods of Hercomite which consists of 85.20 per cent ammonium nitrate, 5.78 per cent sodium nitrate, 3.70 per cent wood pulp, 4.74 per cent apricot meal and .58 per cent ground limestone; and one hod of Herculite which consists of 34.24 per cent ammonium nitrate, 6.52 per cent sulphur, 6.52 per cent sawdust, 8.70 per cent apricot meal, 43.48 per cent sodium nitrate, and .54 per cent ground limestone. The hods of dope when filled remained in the dope house until they were transported to the “mix house.” However, there is controverted evidence as to whether it was normal Hercules practice to let three or four hods of dope accumulate in the dope house. The nearest mix house was about 1,000 feet from the dope house. Yibrogel was being run through the screen to remove impurities when the fire broke out on February 12, 1953. About 50 employees responded to the siren which indicated that there was a fire in the dope house. The dope house exploded, killing 12 employees and wounding 22 others. Only the part of the building containing the dry pipe valve remained intact. The explosion crater was located at approximately the same spot on the other side of the room from the screen, where the filled hods of dope had been standing.

There is much conflicting evidence as to the possible causes of the explosion and fire,- the condition of the dry pipe valve after the accident; the existence and duration of the fire before the explosion; whether the fire was a high flaming one sufficient to reach and fuse the sprinkler heads, or a burrowing penetrating fire within the oxygen supplying dope compounds; or a flash surface fire which would not open the sprinkler heads; or whether the 68 sprinkler heads collected by Hercules after the accident had fused. The evidence also indicated that: there was no way of predicting what mass of the compounds would be critical enough to explode, or how soon a fused sprinkler head would open; that streams of water from the hoses may have a cooling effect on sprinkler heads.

Hercules had no special fire-fighting employees. According to its customary practices the dope house area was a “nonexplosive” area which meant that in case of fire, all employees who could leave their work in other areas of the 2,800-aere plant were to come in response to the siren and help fight the fire. There is some evidence that the fire fighting activities of Hercules employees were voluntary. The general rules and regulations of Hercules, however, besides warning all employees that the occupation was hazardous, stated: “Fire hose and other fire protection apparatus are for use only in ease of fire; they must not be used for any other purpose without a written order from the office. Every employee should learn where the fire fighting apparatus is and how to use it.” The general rules for the dope house stated: “Fire protection: Use of fire equipment should be learned by each employee.” There is controverted evidence as to whether such fire fighting is proper and accepted industry practice. Hercules had several prior dope house fires, two of which began in the processing screen. There had been other known dope house fires in the industry but only one previously known dope house explosion. The respondents contend that the fire before the explosion was not hot enough to fuse the sprinkler heads; that the explosion occurred through the sympathetic detonation of the accumulated hods of dope; that the explosion caused the fire; that the sprinkler system was not called upon to operate; that the system was not designed to cope with explosions, only fires; that the appellant Hercules was eontributorily negligent in treating the dope house as a safety area, and fighting the fire there, and in not letting respondents choose the best type of sprinkler system; that the individual appellants assumed all risks in voluntarily coming to fight the fire in the dope house. The appellant’s theory of the ease is as follows: that the explosion was caused by the fire which would have been extinguished if respondents’ system hqd worked as it should have, and in accordance with respondents’ description of the system as the best and latest method of fire precaution; that respondents knew for what purposes the appellant required the system; that the flames were high enough to fuse the sprinkler system, and that the system had not worked as the dry pipe valve was waterlogged; that the respondents should have warned Hercules of the danger of waterlogging in the valve and provided some device to draw off the water or at least cautioned Hercules to manually draw off the water. As indicated above, the record contains controverted evidence on all of these matters, including voluminous expert testimony.

The appellant’s chief grounds of appeal, which will be taken up in order, are:

(1) The prejudicial error of the lower court in refusing to submit the issue of res ipsa loquitur to the jury. The requirements for the application of res ipsa loquitur in this state were set forth in Ybarra v. Spangard, 25 Cal.2d 486 at p. 489 [ 154 P.2d 687 , 162 A.L.R. 1258 ], as follows: “ (1) The accident must be of a kind which ordinarily does not occur in the absence of someone’s negligence; (2) it must be caused by an agency or instrumentality within the exclusive control of the defendant; (3) it must not have been due to any voluntary action or contribution on the part of the plaintiff.” We have concluded from the evidence that these requirements were met here and therefore the trial court should have given res ipsa loquitur instructions as the jury could have found from the evidence the facts necessary to apply the doctrine.

As to the first requirement, it was said by our Supreme Court in Zentz v. Coca Cola Bottling Co., 39 Cal.2d 436 at page 442 [ 247 P.2d 344 ] : “. . .it must appear, either as' a matter of common experience or from evidence in the case, that the accident is of a type which probably would not happen unless someone was negligent.” The fire and explosion in the instant case under the evidence could well have been due to some fault in the mechanism of the valve and sprinkler system, or it could have occurred because of the explosive quality of the dope itself. (Olson v. Whitthorne & Swan, 203 Cal. 206 at p. 208 [ 263 P. 518 , 58 A.L.R. 129 ].) .An inference of negligence is permissible where it appears that the accident would not have occurred had negli genee not been present. (Zentz v. Coca Cola Bottling Co., 39 Cal.2d 436 [ 247 P.2d 344 ].)

In La Porte v. Houston, 33 Cal.2d 167 [ 199 P.2d 665 ], it was held that there was no error in not instructing the jury on res ipsa loquitur, since it was equally probable that the lurch of the automobile was due to some mechanism of the car for which the defendants were not responsible as that it resulted from any negligent act on the part of the defendants’ mechanic. At page 169 the court said: “Assuming that defendants were in control of the car while the carburetor was being rechecked, the applicability of the doctrine of res ipsa loquitur depends on whether it can be said, in the light of common experience, that the accident was more likely than not the result of their negligence. (Cases cited.) ‘Where no such balance of probabilities in favor of negligence can be found, res ipsa loquitur does not apply. ’ (Prosser on Torts, p. 297.) ” It cannot be-said here that no such probabilities in favor of negligence can be found. The sprinkler system installed by defendants did not work. Thus, even if the choice of the system was entirely at the whim of Hercules, the respondents could not escape liability on that account. There is evidence of a sufficiently hot fire to make the sprinkler system work and that the fire burned for 20 minutes and charred beams in the ceiling directly above the sprinkler system. The jury could conclude that but for some negligence in its design, construction or in its installation the sprinkler system would have worked. The further question of whether if the sprinkler system had worked it would have prevented the explosion is one of proximate cause and not of negligence. Whether the failure of the sprinkler system proximately caused the damage is a question for the jury. (Better Food Mkts. v. American Dist. Telegraph Co. (1953), 40 Cal.2d 179 [ 253 P.2d 10 , 42 A.L.R.2d 580 ].)

A somewhat difficult question arises here as to the presence of the second requirement. The defendant must be shown to have control of the instrumentality. (Gordon v. Aztec Brewing Co., 33 Cal.2d 514 [ 203 P.2d 522 ].) The evidence here indicates that the sprinkler system and firefighting activities were under the control of the Hercules Powder Company. However, among the few items of uncontroverted evidence in the record are: (1) that the respondent manufactured and installed the system and voluntarily undertook its inspection; (2) that respondent’s supervisor Mott installed a new latch 30 days before the fire; (3) that the valve was not touched or tampered with thereafter. The fact that defendant has not immediate control is not required if there is evidence of no improper handling in the meantime. (Zentz v. Coca Cola Bottling Co., 39 Cal.2d 436 [ 247 P.2d 344 ].) The appellants are not required to eliminate every other possibility of control under the circumstances here presented. (Escola v. Coca Cola Bottling Co., 24 Cal.2d 453 [ 150 P.2d 436 ].)

As to the third requirement, the rule is that a plaintiff may rely on the doctrine of res ipsa loquitur even though he has participated in the events leading to the accident if the evidence excludes his conduct as a responsible cause. (Simmons v. Rhodes & Jamieson, Ltd., 46 Cal.2d 190 [ 293 P.2d 26 ] ; Baker v. B. F. Goodrich Co., 115 Cal.App.2d 221 [ 252 P.2d 24 ]; Jesionowski v. Boston & Maine R. Co., 329 U.S. 452 [ 67 S.Ct. 401 , 91 L.Ed. 416 , 169 A.L.R. 947 ].) As indicated above, there is some evidence here, which, if believed by the jury, would exclude the conduct of Hercules as a responsible cause. The issue of Hercules’ contributory negligence is separate from the issue of the respondent’s responsibility.

The position of the individual appellants, is even more compelling than that of appellant Hercules. Their awards under the Workmen’s Compensation Act precluded any action against Hercules; they joined in this suit, pursuant to statute. They properly rely on Hardin v. San Jose City Lines, Inc., 41 Cal.2d 432 [ 260 P.2d 63 ] ; Stanford v. Richmond Chase Co., 43 Cal.2d 287 [ 272 P.2d 764 ] ; Raber v. Tumin, 36 Cal.2d 654 [ 226 P.2d 574 ]; Summers v. Tice, 33 Cal.2d 80 [ 199 P.2d 1 , 5 A.L.R.2d 91 ], to argue that as to them the instrumentality which caused the accident was in the control of either Hercules or Automatic or both, and the accident could have been caused by the negligence of either or both, and that the fact that respondents blame Hercules for the accident does not preclude the application of res ipsa loquitur to the individual appellants, as they are entitled to their own theory of causation and can recover.

In the very recent case of Seneris v. Haas, 45 Cal.2d 811 [ 291 P.2d 915 ], in holding that there was sufficient evidence of the elements of res ipsa to permit the jury to conclude that the doctrine was applicable, the court clearly stated at pages 826, 827: “In Baker v. B. F. Goodrich Co., 115 Cal.App.2d 221, 229 [ 252 P.2d 24 ], it was said: ‘Where the evidence is conflicting or subject to different inferences, it is for the jury, under proper instructions, to determine whether each of the conditions necessary to bring into play the rule of res ipsa loquitur are present. (Roberts v. Bank of America, 97 Cal.App.2d 133, 137 [ 217 P.2d 129 ].) ’ (Citations.) The conclusion that negligence is the most likely explanation of the accident, or injury, is not for the trial court to draw, or to refuse to draw so long as plaintiff has produced sufficient evidence to permit the jury to draw the inference of negligence even though the court itself would not draw that inference; the court must still leave the question to the jury where reasonable men may differ as to the balance of probabilities. (Citation.) The inference of negligence is not required to be an exclusive or compelling one. It is enough that the court cannot say that reasonable men could not draw it (citation). The existence of the conditions upon which the operation of the doctrine is to be predicated is a question of fact and the right of the jury to find those facts must be carefully preserved.” We think it is particularly important in this case, where there is so much confusing and conflicting evidence that the jury determine which of the many possible probabilities or any combination of them, was the most reasonable.

We conclude, therefore, that under the above rules recently expressed on the subject of res ipsa by our Supreme Court, the evidence of the appellants in this case is sufficient to submit to the jury, the applicability of the rule of res ipsa loquitur, and the refusal of the trial court to so instruct, constituted prejudicial error as to all the appellants.

(2) The instructions to the jury that under the facts of the case it could not infer negligence and that specific acts of negligence had to be proved. These instructions were erroneous if the jury found facts which would bring the doctrine of res ipsa loquitur into play. The effect of the doctrine is that when conditions as indicated in a proper res ipsa loquitur instruction are shown to exist, an inference of negligence upon the part of the defendant arises.

(3) The failure of the trial court to instruct that the individual plaintiffs did not assume the risk of the occurrence. As assumption of risk was not urged by the defendants at the trial, there was no prejudicial error in the refusal of this instruction ... Prescott v. Ralphs Grocery Co., 42 Cal.2d 158 [ 265 P.2d 904 ] and Austin v. Riverside Portland Cement Co., 44 Cal.2d 225 [ 282 P.2d 69 ] were proper eases for the application of the defense and the instructions to the jury were erroneous in not properly pointing out the element of knowledge which is a prerequisite of the defense. In Hawk v. City of Newport Beach, 46 Cal.2d 213 at page 218 [ 293 P.2d 48 ], it was held that the defense was not applicable, unless there was voluntary acceptance of a risk and an appreciation of the magnitude of that risk.

(4) The admission of evidence concerning Hercules ’ precautions after the accident in order to impeach the plaintiffs’ witnesses. The appellants are correct in their contentions that evidence of subsequent precautions is not admissible on the issue of negligence. However, in the instant case, the evidence was not admitted for such purpose but only for the limited purpose of impeaching the testimony of Hercules employees and under the record before us was properly admitted. Witness Butler, Hercules’ Work Manager, testified on cross-examination: “I know of nothing, in all my experience in the manufacture of explosives up to the present moment, that would cause me to classify a dope house-to be properly classified as an explosive building,” and in answering a further question as to the safety of the dope house, “I not only consider it satisfactory and safe but it was the best thing I had seen. It had more sprinkler coverage than any other dope house I had seen.” Witness Jones, Hercules Safety Engineer, testified that Hercules believed that the sprinkler system ordered from Automatic would provide adequate protection. Witness Hedrick, Hercules Dynamite Department Supervisor, called by the defendants, under Code of Civil Procedure, section 2055, testified that he did not consider Hercomite an explosive mixture. Witness Conrady, Hercules Assistant Master Mechanic in the maintenance department, in answering whether he had received any instructions to alter the classification of the dope house after the accident, stated, “Not that I recall.” The jury was properly cautioned and instructed as to the limited nature of the evidence of subsequent precautions. Appellants attempt to distinguish Inyo Chemical Co. v. City of Los Angeles, 5 Cal.2d 525 [ 55 P.2d 850 ], on the ground that there the evidence which was impeached was elicited from the witness on direct examination, and not on cross-examination as in the instant case. They rely on Parkin v. Grayson-Owen Co., 157 Cal. 41 [ 106 P. 210 ] and Daggett v. Atchison, T. & S. F. Ry. Co., (Cal.App) 301 P.2d 923 . In the Parkin ease, the questions were held not to be within the scope of proper cross-examination as they dealt with the different subject of driving the horses, while the issue was whether the horses had been properly hitched. In the Daggett ease, the first two witnesses called by the plaintiff were two employees of the defendant, called under Code of Civil Procedure, section 2055. The court held the questions improper as the defense had not yet presented any evidence subject to impeachment.

(5) The use of the United States Army Ordnance Safety Manual to ask questions of plaintiffs’ expert witness, Lewis, and of Cook, defendants’ expert witness. Appellants rely chiefly on Lilley v. Parkinson, 91 Cal. 655 [ 27 P. 1091 ], for the rule that it was improper to question a witness about a work which the witness had not relied upon, as the only effect of such questions was to place before the jury the opinion of the author of the book. Respondents correctly point out that this reason is missing in the instant ease, as the manual had been introduced into evidence and brought to the attention of the jury, when Dr. Cook was questioned. As to the use of the manual on the cross-examination of Lewis, while there are some early cases which indicate that medical texts could be used on cross-examination to impeach a witness who had not relied upon them (Fisher v. Southern Pac. R. R. Co., 89 Cal. 399 [ 26 P. 894 ]; People v. Hooper, 10 Cal.App.2d 332 [ 51 P.2d 1131 ]) we agree with the appellants that the more recent cases (Lewis v. Johnson, 12 Cal.2d 558 [ 86 P.2d 99 ] and Gluckstein v. Lipsett, 93 Cal. App.2d 391 [ 209 P.2d 98 ]) indicate that the text could not be so used. It is a question, of course, whether the rules of the above cases which dealt with medical treatises should also be applied to such documents as the government manual in the instant ease. The medical treatise cases would seem to be an exception to Code of Civil Procedure, section 1936. We are not saying, however, that Code of Civil Procedure, section 1936, is applicable to the government manual in the instant case. In the instant case, the record indicates that the trial court carefully considered the matter, and there was no error in the ruling. The questions from the government manual can be said to have been used only to test the credi bility of the witness, rather than for impeachment purposes. In Parker v. James Granger, Inc., 4 Cal.2d 668 [ 52 P.2d 226 ], the document was not admitted as evidence as in the instant case. Furthermore, Dr. Cook testified that he knew of the government manual.

(6) The striking of witness Frome’s testimony on defendant’s motion. Appellants argue that the striking of Frome’s testimony was prejudicial error as it excluded evidence of industry custom and practice, which is relevant to the negligence of Hercules. However, the issue here was not whether such evidence of industry practice is admissible if relevant—on this point, we agree with the authorities cited by the appellants. The issue here was whether the witness’ experience and knowledge was sufficient to establish industry practice. The trial court said the evidence was too remote and lacking in relevancy. In order to be relevant, the conditions must be the same or similar (65 C.J.S. 1048; Am.Jur. 1015, 1017). Evidence of conduct in a particular situation or in a few instances is not sufficient to establish custom. (Longuy v. La Societe Francaise, 52 Cal.App. 370 [ 198 P. 1011 ] ; McStay v. Citizens etc. Bank, 5 Cal.App.2d 595 [ 43 P.2d 560 ] ; Olsen v. Realty Hotel Corp., 210 F.2d 785 .) Appellants rely particularly on Burke v. John E. Marshall, Inc., 42 Cal.App.2d 195 [ 108 P.2d 738 ] for their argument that witness Frome’s experience was sufficient to establish industry custom. However, the stevedore, whose testimony was admitted in that case, was the plaintiff. As the appellants indicate, this case involved many close questions of fact. We cannot say that the trial court erred in striking the testimony of the witness Frome.

(7) The exclusion of the testimony of witness Krogfoss. The record indicates that the court first sustained objections to the questioning of this witness on the ground that the witness who was a mechanical engineer dealing with valves was not shown to be experienced in the particular type of dry pipe valve involved in the instant case. Moreover, when a proper foundation had been laid, the testimony was admitted. There was, therefore, no error prejudicial to the appellants here.

(8) The individual appellants contend that although the trial court told the jury that the forms of verdict presented to them constituted “every possible verdict,” the forms did not include a verdict against the defendant Automatic alone. When the forms of the verdict were submitted to the jury, the individual appellants made no ohjection. The jury returned to the courtroom after some deliberation as apparently they did not understand that if they should decide in favor of the individual appellants and against the appellant Hercules, they would have to sign two forms of verdict that had been submitted to them. During this discussion, the trial court said: ‘1 There is no problem as to the defendants because they are regarded as one.” The individual appellants made no objection at this time, nor when the verdict was rendered. As the record indicates that the forms of verdict did not include a verdict against Automatic only, the next question is whether the statement of the court regarding the forms of verdict is an instruction within Code of Civil Procedure, section 647, which enumerates matters which are considered excepted to. We think that it is not, and that this matter falls under the well established rule that the appellant’s failure to object to the form of the verdict is a waiver of any defect of form. (Brand v. Norris, 121 Cal. App.2d 367 [ 263 P.2d 456 ]; Joerger v. Pacific Gas & Electric Co., 207 Cal. 8 [ 276 P. 1017 ]; Brown v. Regan, 10 Cal.2d 519 [ 75 P.2d 1063 ].) In Hager v. Gordon, 171 F.2d 90 , cited by the individual appellants, the forms of verdict contained erroneous valuations of the property the plaintiff sought to recover; such instruction interfered with the free exercise of the judgment of the jury. Such was not the case here.

(9) There was no error in the denial of the individual appellants’ motion for a directed verdict.

In view of the prejudicial error in failing to give the instructions covering the doctrine of res ipsa loquitur and the giving of the erroneous instructions that the jury could not infer negligence and that specific acts of negligence had to be proved, the judgment must be reversed.

Judgment reversed.

Dooling, J., and Draper, J., concurred.

A petition for a rehearing was denied June 28, 1957. Respondents’ petition for a hearing by the Supreme Court was denied July 26, 1957, Peters, J., and Wood, J., sitting pro tern, for Traynor, J., and Schauer, J. Shenk, J., and Spence, J., were of the opinion that the petition should be granted.

A hearing was granted by the Supreme Court on December 5, 1956. The final opinion of that court is reported in 48 Cal.2d 655 [ 313 P.2d 557 ].