208 TRICHOBRANCHUS GLACIALIS. 



and one half inserted in the tissues, the shafts dilating a little from the base, then 

 remaining cylindrical till the commencement of the very narrow wings, which have 

 minute striae directed outward and upward, after which they taper to a fine hair-like 

 curved tip. They thus appear to represent the first stage of the development of wings 

 on a bristle. The bristles slope outward and backward in the preparations, but are 

 directed forward in life, the convexity of the terminal curve being in the same direction. 

 The shorter forms often alternate with the longer, and their number corresponds nearly 

 with that of the longer, viz. six in each tuft. There is also a slight gradation in the size 

 of the longer bristles from the dorsal to the ventral edge. Below each bristle tuft is a 

 row of hooks with elongated curved shafts (Plate CXXVII, fig. 4), which increase from 

 the base upward to the shoulder, above which the neck is distinctly narrowed, the head 

 again expanding so as to resemble with the main fang a bird's head. Above the main 

 fang the rounded crown has a series of four to fiye smaller teeth. This kind of hook is 

 characteristic of the bristle-bearing segments. 



A series of vertically flattened uncinigerous lamellae occur on the succeeding segments, 

 and some are broader at the tip than the base. They bear at their apices a row of 

 minute avicular hooks (Plate CXXVII, fig. 4 a), having short, broad basal processes with 

 a convex inferior outline, a posterior outline in which a deep sinus occurs above the 

 basal process, and an anterior outline which in some has a trace of a process beneath the 

 main tooth. The latter is of moderate size, but the teeth above it are proportionally 

 large, so that this hook does not present the same disproportion between the first and 

 succeeding teeth present in the long anterior hooks. In lateral view four to five teeth occur 

 above the chief fang, and in reality they form a rounded crown with their points curved 

 obliquely downward. Malmgren, while noting the distribution of the hooks from front to 

 rear, does not sufficiently define the structure of the posterior hooks. 



Habits. — It has the appearance and habits of a tubicolar species, yet no tube was ever 

 found with it on the West Coast. It is possible that there it lives in a tunnel in the mud. 

 In other localities the tube consists of tough mucus, often with external shreds, or minute 

 grains of sand attached to it. * 



The T. massiliensis of Marion from Marseilles has only three teeth in the posterior 

 hooks, and the rostrate hooks extend to the sixteenth setigerous segment. Possibly this 

 is only a southern variety of T. glacialis. Then, again, the Octobranchus Giardi of Marion 

 and Bobretzky ] would appear to be closely allied to T. glacialis, except for the four pairs 

 of branchiae, which occur as simple processes from the segment bearing the eyes and the 

 three following. The hooks are similar. This form would appear to be an intermediate 

 type, especially in branchiae, between Trichobranchus and Terebellides (Hessle). The Fili- 

 branchus roseus of Malm has two pairs of simple branchiae and seventeen bristled segments. 



De St. Joseph (1894) describes twenty eye-spots on each side, and he found three 

 pairs of segmental organs respectively in segments 3, 4 and 5, the second pair being larger 

 than the others. He could not make out the three posterior pairs mentioned by Meyer. 

 The alimentary canal agreed with that in Ainphitrite. His example, however, was young 

 and incomplete. 



Hessle (1917) distinguishes Trichobranchus glacialis from T. roseus by the fact that 

 1 ' Ann. Sc. nat./ 6 e ser., t. ii, p. 87, pi. x, fig. 21, and pi. ii, fig. 21 a and b. 



