¶1 Independently of the question raised, we are struck with the fact that the bond has been altered. When signed by Bratten, it stood “We, John Fennimore, William Hukill, Abraham Bratten, Thompson Wilson, and William Welden.” It now stands with the last two names erased; and the bond having been in Herdman’s pos *397 session, it must be presumed that he made the alteration, unless he show the contrary. The alteration, even in an immaterial point, avoids the deed. It is not the deed that Bratten signed. 2 Starkie, 476.
¶2 If the plaintiff has any testimony to show that this alteration was not made by his consent, we will hear it.
¶3 But the court also inclined to the opinion, that the other point was against the plaintiff. Not on the point of delivery as an escrow merely, but here it was understood by the parties that all the persons named in the deed should sign it. The case cited from Cranch doesn’t come up to this. The plea is not there non est factum, but a special plea that the deed was delivered as an escrow, which don’t amount to non est factum, and which was not sustained, and could not be, because a deed cannot be delivered to the party as an escrow. Non est factum is a better plea in this action, than per fraudem. The idea of fraud is not sustained by the alteration in this deed, or by the violation of any understanding the parties had. It was no doubt the intention to send the bond to Wilson and Welden; and the sheriff afterwards, considering Bratten and Hukill good enough with Fennirnore, thought he had the right to strike out the names of those who did not sign. We think this vitiates the deed ; and have a strong impression also, on the other point against the plaintiff.
¶4 Judgmént of nonsuit.