after stating the case. His Ho norias we infer, (though the statement as to this part of the case is by no means as clear as it should have been) found as a fact, that the sale was made subject to the homestead as allotted to *98the defendant in the execution, and that the sheriff’s deed was intended to convey only his reversionary interest in the land.
These findings are conclusive alike upon the parties to the action and this court, and being so, we cannot perceive how the case differs from that of Barrett v. Richardson, 76 N. C., 429, in which it was said that where laud is sold at execution sale “ subject to homestead,” the purchaser takes it with the encumbrance — even though the debt be one against which no such homestead right exists, and in the absence of all direct authority on the point, it \y°uld seem that reasoning on general principles must lead to a like conclusion.
A sheriff sells not by virtue of any property in himself, but as the mere instrument of the law, and his deed can convey just that which was actually sold and no more, and the rule that deeds are taken most strongly against the makers does not apply to it, for which see Knight v. Leak, 2 Dev. & Bat., 133, and the other cases referred to in the carefully prepared brief of defendant’s counsel..
It is very true that the sheriff is presumed to sell the whole interest of the defendant in the execution, and when he does so and makes a deed accordingly, it will effectually convey every such interest, even though not within the contemplation, or knowledge of the officer at the time. But when he expressly limits the interest sold, and is understood to do so by persons present, it would be most unjust to bold that a greater interest passed.
No error. Affirmed.