178 DISTRIBUTION. 



excretion from the cocnosarc. It is sometimes fouiicl in the single-rowed graptolites to have 

 become detached from the test or chitinous perisarc, leaving behind it a furrow in which it had lain ; 

 this furrow being in the more perfect state of tiie fossil converted into a tube by a thin extension 

 over it of the test. 



Though the rod would thus form an extremely excejjtional structure, its presence can hardly 

 be regarded as offering an insurmountable obstacle to the admission of the graptolites into 

 immediate relation with the Hydroida. Until lately a similar structure would have quite as 

 justly excluded from the Polyzoa any animal which possessed it. Tlie discovery, however, of the 

 living polyzoal genus HhahilopJeura shows that a rod cpiite like that of the graptolite in all points, 

 except in its not being continued beyond the cell-bearing portion, might be developed in an 

 animal possessing in all other respects a typical polyzoal structure.' 



It is true that the extension of the rod in the fossil beyond the limits of the proper 

 hydrocaulus appears to increase the difficulty of reconciling its presence with the hydroid aifinities 

 of the graptolite. I believe, however, that this is, after all, not so anomalous a fact as at first 

 sight it may appear, and that there is reason to believe that the coenosarc invested by a proper 

 perisarc was originally continued along what now appears as a free extension of the rod. Its 

 distal extension would then correspond to what had been the young growing portion of the 

 graptolite, as yet destitute of denticles, and with its periderm so delicate as to be incapable of 

 preservation in the fossil, so that the thin perisarc has perished along with the soft coenosarc it 

 included, its thicker rod-like portion being the only part preserved. 



This view is borne out by the fact that in the very young stage of the graptolite a distal 

 extension of the body along the rudimental rod, and beyond the incipient denticles, may be 

 noticed ; while it is further confirmed by an observation by Dr. Nicholson," who tells us that in 

 some specimens of JDiphi/raplm pristis he has seen the common canal without denticles continued 

 on each side of the prolonged rod. 



The continuation of the rod beyond the denticle-bearing portion at the proximal end of the 

 graptolite may also have been accompanied by an extension of the ca-nosarc and its enveloping 

 perisarc in this direction, the rod alone remaining in the fossil. To this view an observation of 

 Mr. Carruthers ' gives support, for he has noticed the prolongation of the rod at the proximal end 

 of CUmacoyraptus scalaris frequently invested for a short distance by a sheath. 



If this explanation be accepted, the continuation of the rod as a naked filament beyond the 

 denticle-bearing portion of the graptolite need no longer surprise us. A comparison of the rod 

 to the chitinous spines which bristle over the surface of Hydradinia may also here suggest 

 itself; but these spines are not only invested by a coenosarcal layer, but are permeated by canals 

 which are lined by coenosarc, while in other respects the approximation of the graptolites to 

 Ilydractiina offers too many difficulties to allow of its being attempted. 



The lateral spines often present at the proximal end of the graptolite seem to be of a 

 different nature fr'om that of the rod, and would rather appear to be referable to the same group 

 of structures as the chitinous spines and variously formed processes by which the hydrotlieca; and 

 other parts of the perisarc of living hydroids are not unfrequently ornamented. 



' AUman, " On Rhabdopkura," in ' Quarterly Journal of Micros. Science/ Jan., 1869, p. 57, 

 pi. viii. 



" ' Geological Mag.,' vol. iv, 18G7, p. 261, note. 



' Carruthers, in 'Intellectual Observer' for June, 1867, p. 370. 



