delivered the opinion of the court.
*210Assuming, as we must, that the damages awarded by the District Court resulted from the loss of the cap and plate covering the bilge-pump hole, the question to be determined is whether that loss was occasioned by a peril of the sea or by the condition of that covering as it was when the vessel entered upon her voyage. If through some defect or weakness the plate and cap arid the screws which secured it came off, or if the cap and plate-were so made or so fastened.as to be liable to be knocked off by any ordinary blows from objects washed by the sea across the decks, then the vessel was not seaworthy in that respect, and the loss could not be held to come within the exception of perils of the sea, although the vessel encountered adverse winds and heavy weather. By the charter party it was agreed on the part of the vessel that she should be tight, staunch, strong, and in every way fitted for the voyage, and the rule is well settled that the charterer is bound to see that his vessel is seaworthy and suitable for the service for which she is to be employed, while no obligation to look after the matter rests upon the owner of the cargo. The Northern Belle, 9 Wall. 526; Work v. Leathers, 97 U. S. 379. If there be a defect, although latent and unknown- to the charterer, he is not excused. 3 Kent, *205; Valin, Com. Ord. de la Mar. liv. Ill, tit. Ill, Du Fret; art. XII, vol. 1, 654; Lyon v. Mells, 5 East, 428; Work v. Leathers, supra.
As said, on circuit, by Mr. Justice Gray, in The Caledonia, 43 Fed. Rep. 681, 685: “In every contract for the carriage of goods by sea, unless otherwise expressly stipulated, there is a warranty on the part of the ship owner that the ship is seaworthy at the time of beginning her voyage, and not merely that he does not know her to be unseaworthy, or that he has used his best efforts to make her seaworthy. The warranty is absolute that the ship is, or shall be, in fact seaworthy at that time, and does not depend on his knowledge or ignorance, his care or negligence.” In The Glenfruin, 10 P. D. 103, 108, the same rule is- thus expressed by Butt, J, : “ I have always understood the result of the cases from Lyon v. Mells, 5 East, 429, to Kopitoff v. Wilson, 1 Q. B. D. 377, to be that under his implied warranty of seaworthiness, the ship *211owner contracts, not merely that he will do his best to make the ship reasonably fit, but that she shall really be reasonably fit for the voyage. Had those cases left any doubt in my mind, it would have been set at rest by the observations of some of the peers in the opinions they delivered in the case of Steel v. State Line Steamship Co., 3 App. Cas. 72.”
. Perils of the sea -were excepted by the charter party, but the burden of the proof was on the respondents to show that the vessel was in good condition and suitable for the voyage at its inception, and the exception did not exonerate them from liability for loss or damage from one of those perils to which their negligence, or that of their servants, contributed. Liverpool Steam Co. v. Phenix Ins. Co., 129 U. S. 397, 438. It was for them to show affirmatively the safety of the cap and plate; and that they were carried away by extraordinary contingencies, not reasonably to have been anticipated. We do not understand from the findings that the severity of the weather encountered by the Morrison was anything more than was to be expected upon a voyage, such as this, down that coast and in the winter seáson, or that she was subjected to any greater danger than a vessel so heavily loaded, and with a hard cargo, might have anticipated under the circumstances. The especial peril which seemed at one time to threaten her safety was directly attributable to the water taken aboard through the uncovered bilge-pump hole, which rose from eighteen inches about 5 a.m. to seven feet at about 9 a.m., so that she was necessarily sinking deeper and deeper, while the absorption by the guano added to the dead weight, and increased the danger of her going down.
Among other exceptions, libellant excepted to the refusal to amend one of the findings by adding: “ No spars nor sails were carried away and no repairs were made to the vessel at Savannah beyond nailing a few boards on the starboard side where the bulwarks had been broken, which was done by the crew; and the vessel, after discharging, loaded with a cargo of phosphate rock, which is a much harder cargo to carry than guano, and delivered it safely.” These facts were established in substance by uncontradicted evidence, and as they *212tended to show that the schooner was not injured to any appreciable degree, and therefore that the weather was not of such an extraordinary character as would have damaged the cargo to the extent that it was if the vessel had been seaworthy in the respect under consideration, we think the amendment was material, and should in effect have been made. What happened to the vessel after the loss of the bilge-pump plate throws but little, if any, light upon the situation at the time of the loss; and libellant excepted to parts of the thirteenth finding so far as they involved the inference that certain incidents therein detailed occurred before the cap and plate came off, as unsupported by evidence; and also to that part of the sixteenth finding which stated that it was to be inferred that the plate was knocked out “ subsequently to the time when they wore ship after finding eighteen inches of water in the hold ” on the same ground. But without going into these details or inquiring how far they are open to examination, the significant fact is found that although at half-past four in the morning the pumps sucked, indicating that there was then no water in the well, they did not suck (twenty minutes later, as disclosed by the evidence, nor) again until the hole was discovered and stopped up, when they gained on the water, and after several hours freed the vessel.
In any aspect, the real point in controversy is, did the respondents so far sustain the burden of proof which was upon them as to render the probability that the cap and plate were in good condition and knocked off through extraordinary contingencies so strong as to overcome the inference that they were not in condition to withstand the stress to which on such a voyage it might reasonably have been expected they would have beemsabjected ? If the determination of this question is left in doubt, that doubt must be resolved against them.
The 8th, 11th, 12th, 14th, 15th, and 16th findings were as follows:
“VIII. Such bilge-pump holes are not unusual in vessels constructed in some localities. The plates are generally- considered permanent fixtures, not peculiarly susceptible to deterioration from age. Verdigris sometimes forms around *213brass screws, thus weakening the hold of the wood; but waterways located as this was, well covered up and well painted, are not liable to rot, and their reasonable expectation of sound life is largely in excess of twelve years.
“ If the plates and caps which are generally used to cover such holes are not kept tight and secure, the holes become dangerous; but that mode of covering was generally deemed secure by seafaring men, and seldom, if ever, have any accidents arisen from their use.
“ XI. Before the vessel sailed the cap and plate appeared to be in good order, with no indication of looseness. The examination which was at that time made of them consisted of such inspection as could be given by the eye, and to such an inspection they were from time to time subjected. They were not tested either by unscrewing the cap or the plate, or by tapping the plate with a hammer. Tapping with a hammer or unscrewing the cap might have developed any insecurity (if there were any) in the bilge-pump plate. Immediately after the loss of the port-bilge plate (hereafter described) the mate tested the condition of the similar plate on the poop deck, starboard side, by tapping with a hammer, and found it apparently sound.
“XII. The examination which was made of the cap and plate, as set forth in the Xlth finding (viz., by a survey without the use of special tests, unless there is some appearance of defect) is such as a reasonably prudent master or owner might be expected to give them in order to'determine the seaworthiness of his vessel before beginning a voyage.
“XIY. The wood- to which the plate had been fastened looked white' and sound. From the holes out' of which the screws had come part of the clear wood was itself hauled, the splinters hanging around the edges of the holes,' the holes thus presenting a ragged look. The screw-holes were not smooth nor black nor rusty. The wood of this particular waterway in the vicinity of the plate did not look rotten, and when after arrival at Savannah the temporary plugging referred to in the XYIIth finding was removed, and the hole plugged and covered with sheet lead, the timber into which the plug was *214driven and on which the lead was nailed was found solid, and since that time the covering had not been further repaired nor the timber changed m any way.
“ XY. No marks of violence other than the splintering of the wood- about the screw-holes were visible upon the waterway or upon the adjacent bulwarks or stanchions.
“ XYI. As no one witnessed the removal of the bilge-pump plate, direct evidence of the cause of this mishap is not obtainable. It is, however, to be inferred from the facts proved that it was knocked out by something striking violently against it subsequently to the time when they wore ship after finding eighteen inches of water in the well, which would be between 5 a.m. and 5.30. a.m.”
There was no direct evidence that the plate was knocked out, or, if this were so, that it was by some extraordinary collision ; and while the fourteenth finding tends to support the inference of the sixteenth, it will be observed that the tendency of the fifteenth is to rebut it. If .it appeared that the wood was solid, and the screw-holes splintered, the drawing out of the screws might be imputed to a blow or blows; but, on the other hand, if there were no marks of violence in the. vicinity, since such blow or blows to effect the result, if the cap, plate and waterway were in good condition, must necessarily have been of great violence, it seems almost incredible that no marks thereof appeared on the stanchions and bulwarks on the port side, and that nothing but the cap and plate were carried away. And it is proper to note that no survey of the vessel was had, and that respondents introduced no proof of exact measurements to show the height of the cap above the waterway, or of the perpendicular edge, nor was the duplicate cap on the starboard side produced.
If, however, the vessel had been so inspected as to establish her seaworthiness when she entered upon her voyage, then upon the presumption that that seaworthiness continued the conclusion reached might follow, but we are of opinion that precisely here respondents failed in their case.
From the sixth and seventh findings it appears that the vessel was built in 1873; that the bilge-pump hole had not been *215used for four or five years, if at all; and that the cap and plate were painted over whenever the waterway was painted; and from the findings above quoted that these holes were dangerous unless the caps and plates were kept tight and secure; that the hold of the wood might become weakened by the formation of verdigris about the brass screws; that tapping with a hammer or unscrewing the cap might have developed any insecurity, if there were any; that no such tests were applied ; but that the caps and plates appeared all right to visual observation. But this was not enough to establish the fact of security; and the twelfth finding, that examination by the eye is such as a reasonably prudent master or owner might be expected to give such coverings in order to determine their seaworthiness, does not give it that effect. The obligation rested on the owners to make such inspection as would ascertain that the caps and plates were secure. Their warranty that the vessel was seaworthy in fact “did not depend on their knowledge or ignorance, their care or negligence.” The burden was upon them to show seaworthiness, and if they did not do so, they failed to sustain that burden, even though owners are in the habit of not using the precautions which would demonstrate the fact. In relying upon external appearances in place of known tests, respondents took the risk of their inability to satisfactorily prove the safety of the cap and plate if loss occurred through their displacement.
"We are unwilling, by approving resort to mere conjecture as to the cause of the disappearance of this cap and plate, to relax the important and salutary rule in respect of seaworthiness. The Reeside, 2 Sumner, 567, 574; Douglas v. Scougall, 4 Dow, H. L. 269.
The findings that “ at the time of the contract and lading of cargo and commencement of voyage the vessel was tight, staunch, and strong, and in every way fitted for the contemplated voyage; ” that “ there was no latent defect in the vessel which contributed to the injury to the cargo; ” and that “ the whole of said damage to cargo was caused by a danger of the seas, and was within the exception in charter party and bills of lading,” were findings determined by the interpreta*216tion which the law put upon the circumstances of the transaction as stated in the previous findings, and, as such, open to our revision. Sun Mutual Insurance Co. v. Ocean Insurance Co., 107 U. S. 485; United States v. Pugh, 99 U. S. 265; The Britannia, ante, 130; Gilroy v. Price, App. Cas. (1893) 56, 64.
In our judgment these deductions were incorrect, and the-specific conclusions of law did not follow.
The decree of the Circuit Court is reversed and the cause remanded with a direction to enter a decree for libellants for the amount found due by the District Court with interest a/nd costs.
Mb. Justice Brewer, not having heard the argument, and Mr. Justice White, not being a member of the court when the hearing was had, took no part in the consideration and decision of the case.