viii, D. 6 Light: Morphology of Eudendrium 343 



dranth (Plate I, fig. 1). They are 2- or 3-chambered, very 

 similar to those described by Agassiz ('62) for E. dispar. The 

 width of the gonophore is greater than its thickness, and the 

 division of the distal end when it occurs is in a vertical plane. 



HISTOLOGY 



Perisarc. — The perisarc in this hydroid is a transparent white 

 or pale yellow. It is extremely thin, being from 1.5 to 2.5 /*/* 

 in thickness on the pedicel and only slightly thicker on the 

 hydrorhiza. This seems to be much thinner than the perisarc 

 of the better known species of Eudendrium. In E. parvum 

 Warren ('08), with a hydrocaulus only 0.10 mm. in diameter, 

 the thickness of the perisarc is 4.4 fifi, and in E. angustum Warren 

 ('08), With a hydrocaulus 0.19 mm. in thickness, nearly the same 

 as in the pi-esent species, the perisarc is 6.1 /xfi thick. The thin- 

 ness of the perisarc of E. griffini may account to some extent for 

 the recumbent growth habit of most of its hydrocauli. 



As mentioned before, the perisarc of the proximal end of 

 many of the pedicels and hydrocauli is irregularly wrinkled or 

 annulated. There is also a thickening of the perisarc at the 

 junction of the hydrocauli and hydrorhiza, and a thickening is 

 found on the distal end of the pedicels of those hydranths whose 

 gonophores have begun to relax. 



Near the upper end of the pedicel, the perisarc thins very 

 rapidly. In one specimen, in which the normal thickness of the 

 perisarc was 2.5 fin, this thickness was maintained to within 0.7 

 mm. of the sense groove. At 0.15 mm. below the groove it had 

 thinned to 1.6 fifi, at 0.10 mm. it was 1 //. thick, and at 0.01 mm, 

 it was only 0.8 /* in thickness. At a point opposite the lower 

 limit of the "cambium tissue" of Jickili and Seeliger, usually 

 between 0.15 and 0.2 mm. from the base of the hydranth, the 

 perisarc seems to lose its stiffness, becomes very thin, and is 

 much wrinkled (Plate I, fig. 1). It seems, typically, to end in 

 a groove in the cambium tissue near its upper limit. This 

 groove is by no means as definite as that figured by Warren for 

 E. angustum Warren ('08) and E. parvum Warren which seems 

 to be the equivalent of the groove of the sense ring. I do not find 

 a double perisarc as noted by Warren in E. angustum. 



In sections of the pedicels of some young hydranths, where the 

 perisarc is being formed, a very peculiar condition is found. 

 The ectoderm cells, which are comparatively large, seem to send 

 oflf tube-like continuations of the cell walls, which, twined to- 

 gether, form the first perisarcal covering. These tubes seem to 



