HAMMOND, J.

The plaintiff's theory of the cause of the explosion was that the "sole-laying cement" with which the barrel was filled when bought of the cement manufacturer by the R.H. Long Shoe Company, the shoe manufacturer, was composed of rubber dissolved in naphtha or gasoline; that when the barrel was sold as an empty barrel by the shoe manufacturer to the defendant, some of this solution still remained in the barrel; that when the heated tar was pumped by the plaintiff into the barrel through the bung hole, the naphtha or gasoline became vaporized; that the vapor escaping from the barrel and mixing with the air formed an explosive compound; that some of this compound reached and was ignited by the flame either of the gas jet fifteen feet distant from the barrel, or of the retort fire twenty-two feet distant; that there was a flash back from the flame to the barrel, and that when the flash reached the explosive mixture which was confined in the barrel the explosion occurred.

So much of this theory as respects the existence of the explosive mixture is comparatively free from difficulty upon the evidence, but that part which respects the manner in which the explosive mixture inside the barrel was reached by flame is not so clear upon the evidence. Upon this general theory the presiding judge charged the jury as follows:

"The way the explosion took place is a matter upon which you have heard the testimony and in which there is a great deal of difference in opinion. I think the witnesses agree that the explosion must have taken place in consequence of the application of fire to mixed gases somewhere there at the mouth of the barrel, in the bung hole of the barrel, in the barrel or outside the barrel, but how the fire could have reached that gas from the fires which were then burning in that room is a matter upon which there is a great deal of contest, the witnesses for the defendant claiming that it was impossible scientifically that it should have come either from the retort or from the gas jets. I do not recall the details of the testimony, but one of the witnesses for the plaintiff said it could not have come from the gas jet, and another witness ^fn_ for the plaintiff, while expressing his opinion that it might come from the gas jet, said it would be highly improbable."

While there is considerable difficulty in adopting the plaintiff's theory, still, if the jury believed the plaintiff as to the absence of every other flame than the gas jet and the retort fire, they may have come to the conclusion that this theory in the absence of any more plausible one furnished the most reasonable explanation of the explosion, not only as respects the manner in which the explosive mixture was formed, but also as to that in which it was touched by the flame. We cannot say as matter of law that such a finding was not warranted by the evidence.

It is stoutly maintained by the defendant that there was no evidence of its negligence. In support of this contention it is urged that the accident was a very singular one; that similar barrels had been bought by the defendant from the shoe manufacturer, and that no explosion ever had occurred in barrels so bought by the defendant; indeed, that Prentiss, the defendant's representative, never had heard of such an explosion and had no reason to suspect there was any possibility that such a thing could occur. The plaintiff on the other hand contends that the defendant was negligent in failing to inspect the barrel before placing it in the hands of the plaintiff for use.

Of course the defendant did not insure the safety of the barrel; its duty was, not to furnish an absolutely safe barrel, but simply to use due care to see that it was safe. Its duty was one of due care for the protection of a careful servant. It is doubtless true that there are many kinds of tools and materials which may be placed in the hands of a servant without any inspection whatever. Some may be harmless in their nature and construction, such as a broom or a nail; others may be so complicated that the ordinary purchaser is not expected to examine them but is justified in relying upon a respectable dealer from whom they are bought. In the class first above named there is no need of inspection; in the second class an inspection is impracticable. Shea v. Wellington, 163 Mass. 364. Whether there shall be an inspection and how careful it shall be are matters to be determined by the circumstances.

In the present case the defendant was buying not a new barrel, but a second hand barrer. It was one which was capable of holding fluids; and Jennings, the defendant's representative, may fairly be held to have known that fact. The barrel had contained one substance; he proposed to put into it another. He made no inquiry as to the nature of the substance, although as a man of general information he must have known that naphtha and other explosive materials are kept in barrels constructed water tight like the one in question. The only inspection he made was to see whether there was water in the barrel, but that inspection was made not because water might be dangerous to the servant, but because it was harmful to the tar. It was an inspection not for the protection of the servant, but in the interest of the master.

If inquiry had been made, it would have disclosed that the cement contained an inflammable substance; that by placing tar in the barrel certain conditions likely to exist in the gas works, there danger of an explosion injurious to life and limb. And such an inquiry as to the explosive nature of the cement easily could have been made.

While the case upon this branch is close, we think the jury may well have found that under the circumstances there was a duty on the part of the defendant to make an inspection, and that the injury to the plaintiff was the direct and proximate result of a failure to perform that duty.

It is further contended by the defendant that if there was any duty to inspect, it rested as well upon the plaintiff as upon the defendant. But that position is untenable. The plaintiff had the right to assume, until he saw some indication to the contrary, that whatever inspection was reasonably necessary to see that the barrel could be safely used by him had been made by the defendant, and that the only care he was expected to take was to see that there was no water in the barrel, and this not for his own protection but for that of the tar.

It follows that the ruling requested that on all the evidence the action could not be maintained was properly refused.

The hypothetical question put to the witness Blood was properly admitted. The defendant contends that the statement contained in the question that "two of the barrels had been filled and the third one was being filled," was not in accordance with the evidence. But it is to be observed that the evidence was not closed and the presiding judge could not know what on that point it might be at the close. As stated in Anderson v. Albertstamm, 176 Mass. 87, 91, "The jury are instructed to disregard the answers [to hypothetical questions] unless they find the facts as assumed in the questions; but as it cannot be known in advance what may be the ultimate decision of the jury as to the facts in dispute, the usual practice is to allow counsel in framing a hypothetical question, to assume the existence of such facts and conditions as the jury may have a right to find upon the evidence as it then is, or as there may be fair reason to suppose it may thereafter appear to be; and in determining whether a hypothetical question shall be allowed, the judge in many cases must rely to a great extent upon the good faith of counsel in their statements as to what they expect the evidence will be." It was within the discretion of the presiding judge to admit the question and the answer.

The witness Ordway was properly allowed to state what he had seen as to the liability of the naphtha sometimes to dry so as to leave a small amount in a barrel, and as to conditions of old barrels. The evidence had a tendency to show the action of the naphtha when left in the barrel in small quantities. True, the condition of other old barrels was remote and might well have been excluded, but it was all within the discretion of the presiding judge. The witness was also properly allowed to answer certain questions upon the assumption that two or three barrels had been filled before the one in question, for reasons already given as to the admission of the testimony of Blood.

The records of the Massachusetts General Hospital were properly excluded. The defendant does not contend that they were admissible under the common law, but insists that they are admissible under St. 1905, c. 330. But the records were made before that statute. The first section of the statute imposes upon certain hospitals, including, as we understand, the Massachusetts General Hospital, the duty "to keep records of the cases under their care and the history of the same in books kept for that purpose." The words "such records" in the second section embrace only the records which thereafter shall be kept under the first section. The question is not whether the statute is retroactive, as a rule of evidence or of procedure, in the sense in which those words are used in cases like Stocker v. Foster, 178 Mass. 501, and Woodvine v. Dean, 194 Mass. 40, as contended by the defendant, but rather what kind of records shall be admitted. If the records are those described in the statute, then they are admissible without reference to the time of the trial, but if they are not of the kind described in the statute, then they are not admissible, no matter what may be the time of the trial. The records of the hospital were not those described in the statute, and were therefore inadmissible. This statute has since been amended (St. 1908, c. 269), but the case was tried before the amending statute became operative.

So far as respects the admissibility of the records of the Carney Hospital under St. 1905, c. 330, the same rule applies because these records also were made before it was passed. The defendant insists, however, that the records of this hospital are admissible under the common law. While it is true that the records were not made in accordance with a requirement of law and therefore were not legal records within the meaning of the rule that legal records or copies thereof are generally admissible, still it appears that they were made in the usual course of business by a person in the discharge of a duty, who appears not only as the maker of them but as their custodian. If she had died and her handwriting had been proved, in the absence of any other testimony as to the manner in which they were made up, they would have been admissible. As in the case of Townsend v. Pepperell, 99 Mass. 40, it would have been assumed that the records were of facts known to her. The rule applicable to such records ordinarily is that the entries must be made by a person having personal knowledge of the truthfulness of the statements. This test has been applied by this court in the case of shop books offered to prove delivery of goods, and it has been held that where the clerk who made the entries had no knowledge of the facts the entries are not admissible although the clerk testified that he correctly put down the information he received from the person by whom the delivery was said to be made. Kent v. Garvin, 1 Gray, 148. Miller v. Shay, 145 Mass. 162. It is true that this rule has not been applied with the same strictness to other memoranda. But in substance the general principle is the same. In the leading case of Welsh v. Barrett, 15 Mass. 380, 386, in which a bank messenger's memorandum of a demand and notice made by him in the course of his duty was admitted upon proof of his handwriting, he being dead, the principle was stated in these words: "What a man has said when not under oath may not, in general, be given in evidence when he is dead; because his words may be misconstrued and misrecollected; as well as because it cannot be known that he was under any strong motive to declare the truth. Yet there are well known exceptions to this rule, as in questions concerning pedigree. But what a man has actually done and committed to writing, when under obligation to do the act, it being in the course of the business he has undertaken, and he being dead, there seems to be no danger in submitting to the consideration of the jury." And the rule has been adhered to quite generally except where in the course of the business the clerk making the entry receives his information either orally or in writing from various persons whom he cannot expect to remember and whom it will be impracticable to call. To apply the rule in such a case and to require the evidence of every person in the long line of persons who have had anything to do with the transaction recorded, would be practically impossible, and so as a practical necessity the record is admitted upon the oath of the recorder, if alive, or upon proof of handwriting if he be dead. It is probable that this exception has been carried farther elsewhere than in this State. For a general discussion of the subject see Wigmore on Evidence, § 1530, and cases cited in the notes. In our own State this exception seems to have been recognized in Briggs v. Rafferty, 14 Gray, 525; Adams v. Coulliard, 102 Mass. 167.

In the present case the records were produced by the witness Gabagan. It appeared that the records were made by her, and that she was the proper custodian of them. But it further appeared that she never had any personal knowledge of the facts stated therein; that she received slips of paper from Dr. Painter, the physician, and copied them into the record; and that was all she knew about them. The record was offered as evidence to show that the statements therein made were true. As handed to the witness by the physician they were simply statements of the physician as to what the patient had said to him, or as to the diagnosis made by the physician. The records were comparatively recent. It was not shown that the physician was not living and within the jurisdiction of the court. No necessity was shown, therefore, for the introduction of this hearsay testimony. For aught that appeared there was better evidence. Under these circumstances the reason upon which the general rule was based, namely, that the record should be a record of facts of which the writer had personal knowledge, should be applied. The case is not within the above-mentioned exception to the general rule.

In this connection our attention has been called to the case of Donovan v. Boston Maine Railroad, 158 Mass. 450, the defendant contending that it is like the present case and is authority for the admission of these records. In that case Barker, J., says (p. 453): "No entries were transferred to the despatcher's sheet [in Boston] from the sheet kept at the East Somerville station. As telegraphic messages are read by sound, as well as automatically recorded in symbols, these entries stand upon the same footing as if made from oral statements uttered at the indicated station, and audible in the despatcher's office; or, in view of the symbols in which the manipulation of his instrument by the operator who sends the message makes it visible at the receiving station, the entries are as if made from his signals given at East Somerville and visible in the despatcher's office. These entries are not, therefore, governed by the rule applied in the cases on which the plaintiff relies." After some discussion the ground of the decision is thus stated: "In our opinion, because there is no reasonable possibility that any designed untruth had part in placing upon the train sheet the statements of which it is the vehicle, and all known circumstances concerning it favor its accuracy, and because it was an act rather than a declaration, and was sufficiently identified as genuine, it was competent evidence without the production or proof of the death of the operator who sent the messages; and its entries material to the issue were admissible and proper for the jury to consider, notwithstanding the fact that it was made by the servants of the party by whom it was offered."

The difference between that case and the present is that the telegraphing and reading were regarded practically as one act, and it was more like a record of facts than of record of declarations made by another. The case must stand on its peculiar facts and cannot be regarded as authority for admitting these records. They were properly excluded.

We see no error in giving the twelfth and thirteenth instructions requested by the plaintiff. The thirteenth instruction asked for by the defendant was properly refused. For reasons hereinbefore stated it could not be ruled, as therein requested, that there was no evidence from which the jury were entitled to infer that the defendant or its superintendent knew, or ought to have known, of the possibility of such an explosion.

Exceptions overruled.