The objection to the admissibility of the evidence of Vogdes was properly overruled. He stood entirely indifferent between the parties litigant, or if he had any interest, that interest was against the party who offered him as a witness. If Stephenson failed in the suit against Ware, Vogdes remained responsible to Stephenson, and if he succeeded, Vogdes would be certainly responsible for an equal amount to Ware, and might perhaps be also liable to Ware for the costs of *164this suit. This possible liability for the costs of this suit if it resulted in- a recovery against Ware, might have been an objection, if the witness had been offered by Ware, and Stephenson had objected, but is wholly unavailable in the mouth of Ware.

The demurrer to evidence presents a more serious if not a more difficult question. The assumpsits set out in the declaration are direct and original, not collateral. The necessity of shewing such an assumpsit by the proof is twofold. The proof of a collateral liability for goods sold and delivered to another would not support the declaration, though the undertaking were in writing; and if it would, the undertaking in proof being oral, the statute denies a right of action on it. Does the evidence prove a direct and original, or a collateral undertaking ? In ascertaining the facts proved directly or by inference, we must not be unmindful of the effect of a demurrer to evidence. By it the demurrant allows full credit to the evidence of the demurree, and admits all the facts directly proved by, or that a jury might fairly infer from, the evidence. And in determining the facts inferrible, inferences most favourable to the demurree will be made, in cases in which there is a grave doubt which of two or more inferences shall be deduced. In such cases it would not be sufficient that the mind of the court should incline to the inference favourable to the demurrant, to justify it in making that inference the ground of its judgment. Unless there be a decided preponderance of probability or reason against the inference that might be made in favour of the demurree, such inference ought to be made. The demurrer withdraws from the jury, the proper triers of facts, the consideration of the evidence by which they are to be ascertained ; and the party whose evidence is thus withdrawn from its proper forum is entitled to have it most benignly interpreted by the substitute. He ought to have all the benefit that might have resulted from a de*165cisión of the case by the proper forum. If the facts of • • i • i < the case depend upon circumstantial evidence, or inferences from facts or circumstances in proof, the verdiet of a jury ascertaining these facts would not be set aside, merely because the court might have made inferences different from those made by the jury. To justify the grant of a new trial, when it depends on the correctness of the decision between different inferences to be drawn from the evidence, it would not suffice that in a doubtful case the court would have made a different inference. The preponderance of argument or probability in favour of this different inference should be manifest. When the question is whether or no a fact ought to be taken as established by the evidence, either directly or inferentially, in favour of the demurree, I do not know a juster test than would be furnished by the enquiry, would the court set aside the verdict, had the jury, on the evidence, found the fact ? If the verdict so finding the fact would not be set aside, it ought to be considered as established by the evidence demurred to.

In the case in judgment, the evidence was all parol, and adduced by the plaintiff. In ascertaining the facts established by it, we must look to all of it, and especially in ascertaining the facts established by any one witness, every thing stated by him, as well on his cross examination as on his examination in chief, must be considered. Facts imperfectly stated in answer to one question may be supplied by his answer to another; and when from one statement considered by itself an inference may be deduced, that inference may be strengthened or repelled by the facts disclosed in another. Under the guidance of these principles I proceed to the enquiry, does the evidence demurred to prove, directly or inferentially, a direct and original undertaking on the part of Ware to pay for the goods Stephenson might furnish to Vogdes?

*166To support the affirmative, reliance is mainly placed on the answers of Vogdes to the fifth and sixth questions on his examination in chief. By neither of these answers is an assumpsit of any kind directly proved. Ware stated the relation between himself and Vogdes, from which he was to become the debtor of Vogdes, and Vogdes’s want of supplies from Stephenson’s store, and that he (Ware) could not pay Stephenson until after harvest. An assumpsit on the part of Ware of some kind is fairly inferrible from this statement taken by itself, but what that assumpsit was is not directly stated. Now, that assumpsit might have been a direct and original undertaking by which Ware alone became the debtor for the supplies to Vogdes, or it might be an assumpsit to see Stephenson paid, or it might be an undertaking to retain in his hands, of the money for which he might become the debtor of Vogdes, as much as was necessary to pay the debt that Vogdes might contract with Stephe?ison, and apply it in discharge of the debt. If the facts of the case depended on this isolated view of the evidence, I should, on the principles before stated as applicable to demurrers to evidence, consider it the duty of the court to select, of these implications, that most favourable to the demurree, because.such might not unreasonably be the implication of the jury; though the latter part of the answer to the sixth question would incline me to adopt the last of the above stated implications. But the first of the implications, (which is the only one that would justify a judgment in favour of Stephenson on the demurrer,) is directly confronted not only by facts stated by the witness, but by the strong presumptions which flow from other unquestioned facts. In answer to different questions, he states that he understood Ware was to stand security for him; that Ware was to pay for the goods out of the money that might be due witness when the building was finished ; and if on a settlement no money was due from Ware, he was not to *167pay the account. These facts directly repel the idea that the undertaking of Ware was original and unqualified : and this is corroborated by fair if not necessary implication from the facts, that Stephenson applied to Vogdes to deal with him, that Vogdes was charged on the books of Stephenson with the articles supplied him, and that Stephenson frequently applied to Vogdes for orders on Ware, and obtained one for the amount of his account against Vogdes. The statements in Vogdes's evidence, that he supposed Ware was to be paymaster, and of the witness Lulce of his impression that the account was raised on the credit of Ware, which were relied on in argument by the appellee’s counsel are quite compatible with the supposition that the undertaking was collateral; and so I think it must be regarded by the court.

Whatever doubts may at one time have existed respecting the undertakings within the scope of the first section of the statute of frauds, it has long since been definitively settled that when the undertaking is for a consideration to be received by, or articles to be supplied to, a third person, if the transaction be such that the third person is responsible to the person who supplies the articles, or from whom the consideration proceeds, the undertaking is collateral, and if oral is not binding. Matson &c. v. Wharam, 2 T. R. 80. Cutler v. Hinton, 6 Rand. 509. The doctrine of Matson &c. v. Wharam has never, since it was decided, been effectually questioned in the courts of Westminster. In the case at bar, I think it clear beyond doubt, that whatever might have been the undertaking of Ware, Vogdes was responsible to Stephenson. No one would, I presume, doubt that if Vogdes were the defendant in this action, Stephenson, on the evidence in this case, would be entitled to judgment.

My opinion therefore is that the judgment is erroneous and ought to be reversed, and judgment entered on the demurrer to evidence, for the demurrant.

*168Cabell, J. concurred.

My first impression of this case on the argument was, that the assumpsit to pay for the goods to be delivered to Vogdes, the party who received them, was a direct assumpsit, and not an alternative or collateral assumpsit, W'ithin the statute of frauds; and a very minute examination of the evidence set out in the demurrer has confirmed that impression. Vogdes, to whom the goods were to be delivered, was a housecarpenter and joiner, an entire stranger in the place, and in debt, as appears by the evidence of Lulte, who says he was under the impression that the credit for the said account was given to the defendant and not to Vogdes, because he was a stranger in the neighbourhood, without property there, and because there were notes or bonds of said Vogdes brought there by others from Loudoun and offered for sale at a discount. This is strong evidence that Vogdes was not the party intended to be credited by the plaintiff. But let us see what Vogdes himself says. He says that he was a stranger in the place; that he was engaged in building a house for the defendant, which, it appears, cost 2250 dollars; that the plaintiff and others solicited him to deal with them, (no doubt, on the credit of the money he was to receive for the house from the defendant Ware); that he said he must consult colo. Ware. He then states the conversation that passed between the plaintiff and defendant. He says, colo. Ware stated to Stephenson the plaintiff, that the witness was engaged in building for him, and would want to be furnished with different articles out of his (Stephenson's) store, and that he (Ware) could not pay him for the articles until next harvest come one year; to which Stephenson the plaintiff assented. This, I think, was manifestly not an alternative promise to pay the debt of another, in the words of the statute. Vogdes seems to have been no *169party to it, and of course was not intended to be bound by it. He was standing by, and was to receive the articles. He was not consulted as to the time of payment. The promise was not collateral to any undertaking of his, but, I think, a direct promise by Ware to pay for the articles. In a suit against Vogdes, could he be charged in the terms of the promise made by Ware? I think not. To the sixth question put to him, upon whose promise to pay for the goods in said account were the articles furnished ? his answer was, “ I was induced to believe they were furnished upon colo. Ware's agreement to pay for them.” But I lay no stress upon this, or any other of the opinions of the witness. I rely upon the facts only that he states. He was an ignorant, though, from his testimony, a very honest man. He goes on: “I was a stranger, and did not think mr. Stephenson would credit me, but it was to come out of the money that colo. Ware would owe me at the finishing of the building.” In answer to another question, “ In the conversation you have spoken of, was the understanding of the parties that mr. Ware was to stand as security for you ?” he says, “ That was the way I understood it.” In another place he says, “ Mr. Ware was to be my paymaster, and to receive it out of such money as might be due me on settlement.” All this and much more may be descanted on, but I rely on the terms of the promise made by the defendant to the plaintiff when the latter assented to deliver the articles to Vogdes. There is a great deal of cross examination, but no answer that contradicts the facts stated by him as to the terms of the promise made by the defendant to the plaintiff to pay for the articles to be delivered. The witness’s opinions as to the effect of those facts are not facts, nor contradictory of those facts. As to the orders on Ware which he gave the plaintiff, they prove nothing as to the contract. One of them (which was rejected by the judge) was drawn on *170account of some plank furnished by Vogdes to the defendant. Other orders in favour of the plaintiff on the defendant are spoken of by Vogdes; but these orders, if for money due by the defendant to Vogdes, and if accepted and paid (which is not pretended) would not discharge the assumpsit of the defendant but for their amount, to which the plaintiff could -not object, as every creditor is willing to receive his money even before it is due. Vogdes, in another part of his evidence, says that the plaintiff told him he did not consider the debt to be his, but the debt of Ware the defendant, and refused to take his note for the amount; but the account was kept in the books of the plaintiff in the name of Vogdes. This would be a strong circumstance to prove that the plaintiff held Vogdes bound for the amount, but for the explanation given by Luke, the witness who kept the books. He says, there were other accounts kept in the books, in the names of others for whom Ware the defendant was held solely bound, which were not objected to by the defendant.

In all this evidence there is nothing to prove that Vogdes was expressly or impliedly bound to pay for the articles delivered to him by the plaintiff on the assumpsit of the defendant to pay for them, or to change the character of his assumpsit as stated by Vogdes. The contract of Vogdes was with Ware the defendant to build him a house, and not with the plaintiff. It is a well settled principle that where the party to whom goods are delivered on the promise of a third person to pay for them, is bound to pay for them, the assumpsit is not direct but collateral, and, if not in writing, within the statute of frauds : and the converse of the principle is also settled. In the case before us, I think it very clear that Vogdes was not bound to pay for the articles delivered to him. On the evidence in the demurrer, I think that a jury ought to have found a verdict for the plaintiff, and that no new trial ought to have *171been awarded, if asked for. If the terms of the contract had been less strong, yet on the ground of the utter improbability that, in the circumstances of Vogdes as proved by the evidence, the plaintiff would have given him any credit, I think such a verdict ought not to have been set aside. I have examined many of the cases, but shall only quote one, to prove that the assumpsit of Ware the defendant was a direct assumpsit, and not a collateral promise within the statute of frauds. It is the case of Dixon and others assignees of Moore v. Hatfield, an action of assumpsit on the following agreement: “I Richard Hatfield do agree to pay mr. J. Moore £50. for timber &c. out of the money that I have to pay Wm. West, provided West's work is completed.” At the trial before Best, chief justice, it appeared that West had undertaken, for a certain sum, to complete the carpenter’s work on the house in question, and find all the materials, but being delayed for want of credit or funds to procure timber, it was supplied by Moore, on the defendant’s signing the above agreement. The jury found a verdict for £50. and. Vaughan, Serjeant, moved for a rule nisi to set it aside, on the ground (among other objections) that the agreement was in effect a guarantee to pay Moore in the event that West failed to pay him, and if it was a guarantee, the consideration for the defendant’s undertaking was not sufficiently expressed. But the court were clear that this was not a collateral but a direct undertaking, and the rule was refused. It is true, the agreement in this case was in writing; but I quote it to shew the character of the assumpsit of the defendant in the case before us. The circumstances of the two cases are very similar. The person for whom the agreement in the case cited was made, wanted credit:- so in the case before us. In England, the statute of frauds requires a consideration though the promise be in writing; under our statute, all that is required is that the promise should be direct *172and not collateral, not a guarantee, and then no writing is necessary. Compare the words of the agreement in Dixon and others v. Hatfield, with the words used by Ware the defendant and addressed to the plaintiff, as proved by Vogdes, and I think it impossible to doubt that his assumpsit to pay for the articles to be delivered by the plaintiff to Vogdes was a direct and not a collateral promise.

On the whole, I think the judgment of the court below was correct and ought to be affirmed.

I concur in reversing the judgment, and entering judgment for the defendant on the demurrer to evidence.

Judgment of circuit court reversed, and judgment entered for the demurrant to evidence.