260 BE. G. H. FOWLEE — BISCAYAN PLANKTON : 



the successive stages seemed to overlap as in rotuudata (p. 273). Wliilc this overlap 

 may be jiartly duo to the comparatively rough method of measurement adopted, it is 

 certainly the case at one point that the highest measurements of one stage coincide with 

 or actually overlap the lowest measurements of the next oldest stage. Thus, for 

 example, in a few male specimens measured to a second place of decimals, at Stage II. 

 (lower sexual stage) were found the lengths 0-82, 080, 0"78, 0'75, 0'73 mm. ; and at 

 Stage III. (without secondary sexual characters) the maximum length observed 

 was 082 mm. 



On the whole, the possible errors were so great that it did not seem worth while to 

 take a long series of measurements, as has been done with most of the other species 

 described. There are not only the personal error, the errors due to parallax and 

 refraction, the error due to faulty position of the specimens, and the large error due to 

 distortion of the shell, to be considered ; these all apply also to bigger species (in which 

 they would seriously affect the second place of decimals) ; but, further, all these errors 

 are intensified by the minuteness of the specimens of curia, so that they together amount 

 to a far larger " percentage error " than in the larger forms. 



A good many specimens were measured to a second place of decimals (about -/o nim.). 

 The results seemed to be perfectly worthless. TJie only chance of accurate measure- 

 ment to a second place was of an undistorted shell from the inside — that is, of the riglit 

 shell when the left has been turned back ; this was often impossible when the animal's 

 body lay over the posterior border, or its appendages over the anterior border ; and a 

 comparison of such measurements with tliose taken before the shell was opened (when 

 the true edges cannot be seen), or taken from the turned-back shell (distorted by 

 opening), gave widely difi'erent results. 



If my association of digmatica with curta be correct, we can recognize in the males a 

 Stage I. {^' slUjmatica stage"), with a mean somewhere about 10 mm.; a Stage II. 

 {"•curta stage "), which may be at least as long as 0'82 mm., and as short as 0"73 mm. ; 

 and a Stage III. (without secondary sexual characters), which may be as long as 

 082 mm., and is probably a later stage than the numerous specimens of about 0"55 mm. 

 In tlie female the stages seem to be impossible to separate with any precision, as there 

 is less morphological alteration from stage to stage than in the male; but I think that 

 Stage I. had a mean of about 09 * or 1"0 mm., Stage II. about 07 mm., and Stage III. 

 about 0"5 mm. 



This total result is eminently unsatisfactory as compared with that obtained from the 

 larger species ; but a longer series of measurements from the ' E<esearch ' collections would 

 not better it, because the weakest point of fewest measurements lies about Stages I. 

 and II., and the larger specimens (of 0*9 and more mm.) were entirely confined to the 

 lower strata (below 400 fathoms), from which every available specimen has already been 

 measured. Nor would additional measurements of the smaller specimens (except in 

 order to draw a line between male Stages II. and III.) avail to resolve the overlapping 



* It is noticeable that accordiug to MiiUcr the curta group is almost unique iu having the males equal to or 

 longer thau the females. 



