726 Bashford Dean tAemorial Volume 



drawing. I note that when the albumen is removed close to the egg (so that the surface may be 

 better examined) the germinal area passes out of sight below the equator of the egg. Indeed 

 it is quite probable that the germinal area has its normal position nearer the upper pole before 

 the tension of the albumen is relaxed by the rupture of the wall of the capsule. In these 

 stages the earliest [Figure 1, plate I] has the fewest furrows. All show the furrows clustered 

 in the upper part of the egg, and extending thence more or less radially toward the periphery. 

 In side \dew [Figures 3 to 5, plate I] (the egg ha\'ing been rotated into this position by means 

 of hooked needles thrust into the albumen) the furrows are seen to pass down the sides of the 

 egg in nearly parallel series precisely as they do in Lepidosteus, Amia or Jvjeaurus: some of the 

 furrows extend lower than their fellows, and all round out, flattening at the ends. [Figure 6, 

 plate I, shows a somewhat obHque \dew of the lower hemisphere, with the furrows converging 

 toward the vegetal pole] . The similarity of such a stage to the blastula of a ganoid is made the 

 more striking by the range of color in the Cestraciont egg. The animal pole is of a pale-straw 

 tone, the lower hemisphere is greenish-yellow and the intermediate (equatorial) zone has usually 

 an orange or brownish cast. [In his notebook Dean states that the equatorial zone has a green- 

 ish color. For the upper hemisphere, the colors are shown in Figure 79, plate VII] . 



Another regard in which the furrows indicate their homology with cleavage lines is 

 their behavior with respect to the downgrowing blastoderm [a product of the germinal disc] . 

 This begins at the side of the segmented "animal pole" of the egg, extends across it and en- 

 closes the egg in such a way that the [yolk] blastopore closes at nearly the opposite point on 

 the equator of the egg to the one where the germ was situated, in this regard suggesting the 

 conditions of Lepidosteus or Amia. In this connection it is to be noted that when the [yolk] 

 blastopore in Cestradon [Heterodontus] is closing one may see through it a few long furrows 

 which belong to that portion of the egg (near the equator) where the Hnes become nearly 

 parallel (Figures 52 to 56, plate V). 



It is none the less an extraordinary thing to maintain that a shark's egg, especially one as 

 large as [that of] Heterodontus, possesses a form of holoblastic cleavage. Accordingly it 

 would not be amiss to consider the objections which might be urged against such a thesis. 

 Let us tabulate the objections as follows, and set against them the facts favorable to the view 

 that a rudimentary holoblastism is present. 



Concerning the Homology of the Furrows of the Egg of Heterodontus 

 to Cleavage Furrows of a Holoblastic Egg: 



■^oto* 



They may be surface wrivik^s ordy. — In this event we might expect them to occur in the 

 mature egg, to be more or less inconstant, and to be subject to change by artificial means. 

 They are, however, absent in the egg about the time of fertilization. They are constant in all 

 early stages examined (about a hundred specimens). They are not altered in shape and 

 position by artificial means, such as pressure or tension; nor do they become obUterated 

 (according to observations on an opened egg which was kept aHve for thirty hours). They 

 can be distinguished after the egg has broken, and it can then be seen that the furrows are not 

 superficial merely, but that they pass deep into the yolk, by actual measurement at least 1.5 

 mm. in the upper part of the egg. Moreover the furrows do not occur at hazard. One always 

 finds a central series of segments in the upper part of the egg, and in the lower part a peripheral 

 series, with furrows nearly parallel; occasionally, moreover, as in the egg of Lepidosteus or 

 Amia, several of the marginal furrows may be traced into the region of the vegetal pole and 

 may even traverse it [Figure 6, plate I] . These numerous, close and constant correspondences 

 can hardly, therefore, be without homological significance. 



