The spermatogenesis of Peripatus (Peripatopsis) balfouri. 313 



be determined ou comparatively thick sections, in which the whole 

 body lies in the plane of the section ; then it may be constated that 

 it is a cup shape in apparently all cases , though the sides of 

 the cup sometimes bend inwards, sometimes outwards, sometimes 

 are parallel, and though the base of the cup may be rounded, pointed, 

 or flattened (very usual when it is apposed closely to the nuclear 

 surface). Thin sections on the contrary show only portions of it, and 

 since the axis of the cup makes all possible angles with the axis of 

 the cell, such sectioned portions may appear ring-shaped, U- or V- 

 shaped, or have the form of two parallel rods (Figs. 101, 108, 109, 

 111, Plate 20; Figs. 116, 118, 119, 120, 122, 123, Plate 21). The 

 possibility comes to mind that the cup shape might be produced by 

 1) the idiozome substance grouping itself peripherally, thereby pro- 

 ducing a hollow sphere, and then 2) by a portion of the wall of this 

 sphere becoming invaginated so as to produce a double-layered cup 

 (as in an invaginated gastrula form). But evidently such a process 

 does not take place, for no stages were seen which could be con- 

 strued as the beginnings of such an invagination, nor yet in any case 

 was the wall of the cup found to be double-layered through it often 

 varies in thickness and density at different points. 



This idiozome body lies always in the cytoplasm at the distal 

 pole of the nucleus (Figs. 77, 80, 94—96, 99, 101, 103, 104, 106, 

 108—111, Plate 20; Figs. 128—134, Plate 21). Frequently it is in 

 contact with the nuclear surface, but is more or less widely separated 

 from the latter in elongated cells, so that it may be said to occupy 

 approximately the centre of the cell body, in so far as this is not 

 occupied by the nucleus. The opening of the cup is directed towards 

 the nucleus, or away from it, or to one side; in the Figs. 112 — 120, 

 122 — 124 the outline of the nuclear membrane (N.Mb) has been 

 shown in order to demonstrate these relations. I could not determine 

 whether these differences in position were due to revolutions of the 

 cup on its own axis, and whether then the different positions could 

 be indications of successive stages in its history; but I think it prob- 

 able that no such revolution occurs, for the cytoplasm extends into 

 the cavity of the cup (as was shown in one case where an abnormally 

 large yolk globule had penetrated into this cavity. Fig. 116), which 

 would probably tend to hold the cup in position. 



In the height of the rest stage two small granules of equal size 

 are to be sometimes seen in the cavity of the idiozome body (Figs. 96, 

 101, Plate 20; Figs. 120, 125, Plate 21), and at first I assumed these 



