582 Profs. A. Dendy and J. W. Nicholson. Influence of 



does not quite coincide with the thickest part of the spicule. The distance 

 of this other whorl (the subsidiary whorl) from the end is 20'5 mm., so that 

 the observed ratio of distance to length is 



20-5/74 = 0-277 



against the theoretical value 0291. The error in the ratio is less than 5 per 

 cent. The theoretical distance of the whorl from the end is 2T5 mm., and 

 the discrepancy is only 1 mm. The displacement is in the same direction as 

 that of the median whorl, apparently a usual phenomenon when the latter is 

 displaced, even though only slightly. 



Fig. 6. — This spicule is also very near the critical stage. The lengths on 

 the two sides of the primary central thickening (which is well marked) are 

 40"5 and 41 mm., the latter being on the side on which the subsidiary whorl 

 is situated The median whorl is nearly 1 mm. from the primary central 

 thickening, but almost at the geometrical centre, and the subsidiary whorl is 

 26 - 5 mm. from the end. The ratio of distance to length is 



26-5/81-5 = 0-325 



-as determined by observation. The spicule is very nearly conical. Typical 

 breadths are 2*1 mm. and 1 mm., at 27'0 and 10 - mm. respectively from 

 the end. The value of n calculated from these is 4/5, and the interpolation 

 formula yields the ratio - 314, differing from the observed value by less than 

 4 per cent. The theoretical distance of the subsidiary whorl from the end is 

 therefore 0-314 x 8T5 = 25'5 mm., so that the discrepancy between theory 

 and observation is 1 mm., and as the spicule is slightly curved, so that the 

 theoretical value is really higher, the result may be even more accurate. 



Fig. 7. — In this case the median whorl practically coincides with the 

 primary central thickening. The whorls are relatively larger, indicating a 

 •slightly more advanced stage of development. 



The length of this specimen is again 8T5 mm., and the subsidiary whorl is 

 .at 27'5 mm. from the end, giving a ratio - 337. Typical breadths are - 5, 

 1-8, 2 - 5 mm., at distances 7*0, 27'0, 38 - mm. from the end respectively, the 

 resulting value of n being 0*937, so that the spicule is almost precisely 

 conical. Calculation gives the theoretical ratio 0"322, again differing from 

 the observed value by less than 4 per cent., while the theoretical distance of 

 the subsidiary whorl from the end is 26 - 3 mm., giving an error of only 

 1*2 mm. in the actual distance. This would probably be eliminated if correc- 

 tion were made for curvature. 



Fig. 8. — The median whorl and the organic centre of the spicule again 

 coincide very elosely, though one end of the spicule has apparently grown a 

 little more than the other, the semi-lengths, measured from the thickest part, 



