Denys also states that the roots of "fir," split into three or four parts, were used in sewing. He apparently used "fir" as a general name for an evergreen. It is probable that the roots used were of the black spruce. The technique of building he describes is about the same as that outlined in the last chapter. He says that the gunwales were round and that seven beech thwarts were employed, practices that differ from those in more recent Micmac canoe building, and he notes the goring of the bark cover. Denys states the paddles were made of beech (instead of maple as was perhaps the case) with blades about 6 inches wide and their length that of an arm (about 27 inches), with the handle a little longer than the blade. He also says that four, five, or six paddlers might be aboard a canoe and that a sail was often used. "Formerly of bark," the sail was made of a well-dressed hide of a young moose. Since it could carry eight or ten persons, the canoe Denys is referring to is obviously a large one. In his building description he does not mention headboards, rail caps, or the end forms. It may be assumed that he was then describing a canoe he had seen during construction but whose building he did not follow step by step. De la Poterie, in his book published in 1722, gives a profile and top view of what must have been a Micmac canoe. The probable length indicated must have been about 22 English feet overall and about 32 inches extreme beam; seven thwarts are shown. Late in the 19th century there appears to have been some fusion of Micmac and Malecite methods of construction, as Malecite built to Micmac forms and vice versa. This apparently did not produce a hybrid form so far as appearance was concerned but it did affect construction, in that inner end-frames were used and other details of the Micmac design were altered. The Micmac, having early come into close contact with the Europeans, were among the first Indians to employ nails in the construction of bark canoes, and this resulted in an early decadence in their building methods. Hence, some examples of their canoes show what the Indians termed broken gunwales, in which the ends of the thwarts were not tenoned into the gunwales, but rather were let flush into the top by use of a dovetail cut or, less securely, by a rectangular recess across the gunwale, and were held in place with a nail through the thwart end and the gunwale member.