38 and 47 were joined together with a paper patch, thought this is now partially detatched. 27 appears to be glued into the binding.
The smooth side of each folio is outlined in blue, and the rough in red.
Based on the collation alone and the contents of the folia, it might be concluded that folia 27 and 38 were originally a single bifolium which formed the outer layer of quire Ⅳ. However, an examination of the parchment shows that both folia are smooth on the front and rough on the back, meaning that they couldn't have been attached together unless one of them were backwards, but the sequence of contents on each one doesn't allow that.
The most likely explanation, therefore, is that each folio is a remnant of a bifolium whose other side is lost, which likely means they formed the outside layers of quires Ⅲ and Ⅴ (which also brings them to the same size as quire Ⅳ). Folio 47, currently attached to folio 38 with a paper hinge, is also missing its counterpart. All three of these missing leaves were likely blank, since there are no gaps in the manuscript's contents at these points (compared to the other manuscripts).
For this visualization, I also combined quires Ⅰ and Ⅱ to create a single quire with a blank first recto side, as is common in illuminated manuscripts. The verso side would typically contain a frontispiece, though this manuscript has none.