20 Edmund B. Wilson 



— tliat is directly over the partition wbich separates the latter — and 

 in cross-sections we find that the young biids overlap, to some extent, 

 the older buds, so as partially to include them within the branch. If we 

 follow out in Imagination the farther growth of the colony, we see that 

 sooner or later the older polyps must become enclosed within the branch 

 by the younger ones lying outside them. Hence each polyp is at the 

 sides and within the branch (i. e. the side turned towards the axis) ad- 

 jacent to two older polyps. The canal-system , by which the younger 

 and the older polyps communicate, lies partly at each side of the for- 

 mer, and partly in the partition between the two older polyps, opening 

 into the young polyp along the sides and along the middle dorsal line 

 as shown in fig. 4. A little consideration will show that if the office of 

 the ectodermic bands be to conduct the nutritive fluids derived from 

 older polyps to the upper part of the young bud , then their most ad- 

 vautageous position is on the dorsal septa , since in the typical arran- 

 gement, as shown in fig. 4, each dorsal septum lies between two rows 

 of openings by which the nutritive fluids make their entrance. This 

 explanation will apply also to other compact colonies. 



If, as seems probable, a form like Veretillum has given rise to the 

 penniform Pennaiulida through forms resembling Kophohelemnon^ then 

 the position of these filameuts in the Penniformes is probably due to 

 simple inheritance. We can perhaps explain the fact that in some of 

 the Penniformes the dorsal filaments are extremely unequal. In fig. 14 

 I bave diagrammatically represented six polyps from a leaf of Hali- 

 sceptrum, as described by Kölliker. Each polyp has one long and one 

 short filament. The six polyps fall obviously into two groups of three, 

 in each of which the oldest is at the riglit. In some cases these groups 

 consist of as many as seven or eight polyps. The polyp-cavities open 

 directly into one another by perforations in the partition-walls between 

 them, as represented by the brokcn lines in the diagrani. 



The important point to observe is that in each polyp the filaments 

 on the right side exteud along the whole length of the partition which 

 separates it from an older polyp; and the advantage of this arrange- 

 ment is obvious when we consider the direction of the currents. This 

 arrangement is probably to be regarded as a secondary one specially 

 acquired by some of the Penniformes. 



IV. Pliylogeny of the mesenterial filaments. 



The late appearauce of the dorsal filaments in the egg-embryology 



