Do substances have formal parts?

Hylomorphism is the Aristotelian theory according to which substances are composed of matter and form. If a house is a substance, then its matter would be a collection of bricks and timbers, and its form is something like the structure of those bricks and timbers. It is widely agreed that matter bears a mereological relationship to substance; the bricks and timbers are parts of the house. But with form things are more controversial. Is the structure of the bricks and timbers best conceived as a part of the house, or is it related to the house in some non-mereological fashion? Kathrin Koslicki argues that substances have formal parts and that forms are best conceived as bearing a mereological relation to substances. This paper shows that her argument fails, given the traditional and plausible distinction between substances and accidental unities. I close with a brief suggestion for a non-mereological construal of forms.