HYDROIDS OF THE HAWAIIAN ISLANDS. 
937 
This. species being a representative of a new family, it seemed advisable to investigate it somewhat 
in detail, especially regarding certain histological features, some of which are worthy of further mention. 
The stem . — Although presenting every appearance of being polysiphonic, this structure is not 
homologous with the ordinary polysiphonic stem. A cross section reveals the fact that we have here 
to deal not with a series of parallel tubes, but with a series of irregular lacunge greatly lengthened 
along the axis of the stem and inclosed in a common matrix, as it were, of chitin. In cross section 
these lacunge are of various sizes and shapes, and on the surface they reveal themselves as cross sections 
of deep irregular grooves, 'rather than tubes. (See fig. 4. ) 
Toward the interior of the section the coenosarc appears often to be wanting in these spaces, but 
toward the surface the lacunae are filled with it. The superficial grooves are filled, and here the 
coenosarc overflows, as it were, and forms a complete investment of the stem. It will thus be seen 
that the coenosarc is much deeper over the superficial grooves than between them, where the superfi- 
cial layer is very scant. Being thus molded into the grooves the coenosarc here forms thickened 
strands, several of which unite to form the hydranth body. 
From an examination of such a section as is shown in figure 4 it seems that the mode of growth 
of the stem is somewhat as follows: Reasoning from homology, it is altogether probable that the 
chitinous parts of the hydrocaulus are formed by the ectoderm, although I have been unable to work 
out the beginning of this process. The deposition of chitin, then, takes place at the periphery of the 
lacunae, the latter thus becoming smaller as the chitin invades them, in some cases, doubtless, being 
almost obliterated. Near the surface the ridges appear to be built up more rapidly than the bottoms 
of the grooves, and the outer or superficial sides of the grooves thus tend to meet or be bridged over 
with the rapidly forming chitin. Thus the spaces which were originally cross sections of deep 
grooves finally become round in section and appear to be sections of tubes, and we have, as a result, 
a stem, which, although morphologically polysiphonic or fascicled, is not so in strict homology, on 
account of the great difference in the manner of its formation. 
Unless one has studied the formation of the skeletons in other ccelenterates, such as corals and 
Hydrocorallinse, it is difficult to understand the mutual relations of endoderm, ectoderm, and chitin 
in this new form. Without entering into further discussion, however, it will suffice to say that the 
relations here and in the Hydrocorallinae are, in my opinion, strictly homologous^ 
The Hydranths . — Cross sections and longitudinal sections of the nutritive “ persons” of the colony 
show that they are typical hydroids, although they bear a distinct superficial resemblance to actinoid 
polyps. They are made up of the ordinary histological layers, the body cavity being simple and 
showing no trace of oesophageal tube or mesenterial filaments. There is practically no proboscis, the' 
mouth being in the center of a flat oral disc, as in the actinians. At its base the hydranth body, or 
rather its foot, becomes continuous with the strands of coenosarc which fill several of the adjacent 
superficial grooves of the hydrocaulus, and it is altogether probable that the lumen of the body is 
directly continuous with the central cavities of these tubes, although the condition of the material did 
not allow of definitely proving this point. The tentacles are solid, noncapitate, and arranged in a 
single circlet around the edge of the oral disc. 
■ Gonosome . — This is the most remarkable feature of the form under consideration, being, in my 
opinion, the most primitive known among the Hydroida. Upon the body of an unmodified hydranth, 
below its middle, is borne a huge, hernia-like protuberance that is sometimes almost as large as the 
hydranth from which it springs. But a single one of these is borne on a single hydranth. Upon 
sectioning this strange gonophore it is found to be a simple sack, opening widely into the body cavity 
of the hydranth and consisting simply of the ordinary layers of the hydranth body, endoderm, 
ectoderm, and supporting layer, or “Stutzlamelle.” 
The male and female gonophores are borne on the same colony and are externally quite similar in 
appearance. Internally, they differ only in the sexual elements inclosed. If the gonophore is female, 
the developing ova are seen to be embedded in the endoderm, the older ones being distal and the 
younger being proximal in position. There is not a trace of any medusoid structure to be seen, care- 
fully prepared sections, both transverse and longitudinal, showing no sign of radial canals or of 
blastostyle or manubrium. In the section figured in figure 5, plate i, the line of ova can be traced 
around the bend where the gonophore joins the hydranth body until the smaller ones are seen 
embedded in the endoderm of the body wall itself. It seems likely, therefore, that they are originally 
differentiated in the endoderm of the body wall and afterwards are carried along with the portion of 
the wall that is pushed outward to form the hernia-like gonophore. Or they may migrate outward 
