'34 



SCIENCE-GOSSIP. 



Moreton luUa goes up to 85 and fasciata down to 

 11 per cent. For details see Table I. 



We have next to consider the question of the 

 presence or absence of bands. Altogether, in 

 T. nemoralis, about 13 per cent, are unicolorous ; 

 in T. hortensis, about 42 per cent. Dormington 

 Quarries have a high percentage (28) of uni- 

 colorous nemomlis ; the De Boinville collection a 

 low one (19) of unhanded hortmsis. Unhanded 

 libellnla are uncommon (5 per cent.), and apparently 

 commonest at Dormington (18 per cent.) ; in 

 rubella they are commoner (12 per cent.). Banded 

 specimens are, as a rule, rare in castanea, but in 

 the De Boinville shells are S2 against iS per 

 cent, banded. For details see Table I. 



It is considered that the normal number of bands 



collection, so these castanea have more numerous 

 bands. As each shell has a possible normal 

 maximum of five bands, we may calculate what 

 percentage of this possible total is present in each 

 form. In all nemomlis it is 70 per cent., low in 

 castanea(ig), highest in libellnla (78) ; though, if mista 

 be reckoned as a separate form, it reaches a still 

 higher average, as many rubella have only 20 

 per cent, present (B. F. 00300). For details see 

 Table II. 



The bands are of very unequal width : e.g., band 

 3 in T. nemoralis var. rubella, B. F. 00300, varies 

 from 0'5, or less, to 4 mm. in width. Hence one 

 should take into consideration the strength as 

 well as the presence of the bands. I have tried, but 

 in vain, to devise some accurate but at the same 



Table II., showing the percentage of banded specimens in which each band is present in 

 the several colodr-forms of tachea nemoralis. the percentage of total possible 

 bands present is calculated on all, not only banded, specimens. 



Locality. 



T. nemoralis var. castanea . . Band 1 



2 



3 



4 



Percentage of possible bands present 



T. nemoralis var. libellnla . . Band 1 



2 



3 



4 

 5 

 Percentage of possible bands . . . . . 



T. nemoralis vars. mista and rubella Band 1 



2 

 3 

 4 

 5 



Percentage of possible bands 



All T. nemoralis . . . . Band 1 



3 

 4 

 5 

 Percentage of possible bands 



De Eoinville. 

 11-76 

 27'45 

 S4-3I 

 6667 

 5&S6 

 4060 



500 



54-89 

 9239 

 S3-69 

 82-61 

 67-60 



42-10 

 46-99 

 9398 

 7S-57 

 7669 



62 2S 



4I-92 

 4790 

 92-41 

 79-24 

 76-85 

 61-70 



Moreton. 

 No bands. 



81-29 

 81-29 

 100 

 90-65 

 9065 

 SS78 



79-83 

 84-67 

 100 

 S 9 -5S 

 89-23 

 8S-33 

 80-28 



83-57 

 100 

 S9-91 

 80-67 

 83-62 



Hunderton. 







5° 

 IOO 



50 

 50 

 16-7 



7I-43 

 7 J -43 

 100 



7S-57 

 7857 

 7S57 

 5821 

 6S-66 

 97-01 

 71-64 

 7164 

 6649 



5904 

 69S8 



97 - 59 

 7229 

 72-29 

 6553 



Average. 



3-92 

 25-82 

 61-44 

 3S-89 

 35-62 

 19-1 



67-57 

 6920 



97-47 

 8430 



S394 

 78-32 



6005 

 66-77 

 97-00 



79-93 

 7930 



72-37 

 60-41 

 67-12 

 9667 

 S048 

 79-60 

 70-28 



for each specimen in Tachea is five (hence, with 

 Cryptomphalus, the atrocity of " Pentetaenia "). 

 I have tabulated the occurrence of these bands in 

 the various colour-forms of T. nemoralis from three 

 localities. We find that in all these specimens, 

 band 1 occurs in 60 per cent, of the banded (not 

 total) specimens, 2 in 67, 3 in 97, 4 in 80-5, 5 in So. 

 Thus the order of most frequent occurrence is 

 3, 4, 5, 2, 1. Band 3 is very seldom absent when 

 any other bands are present, as in libellnla from 

 Moreton and Hunderton, and in rubella and mista 

 from Moreton. The De Boinville shells have 

 several peculiarities ; all bands are less frequently 

 represented, and 1 and 2 are present in less than 

 half the banded specimens ; on the other hand, 

 just as banded castanea are numerous in this 



time practicable method of measuring the width of 

 the bands ; the curvature of the shell, which is not 

 equal all round, and other practical details, have 

 at present baffled me. Measurement under the 

 microscope on a travelling stage seems the best, 

 but there are obvious difficulties, and the process 

 is a very long one, and it is a most essential and 

 vital point in this work, where hundreds of speci- 

 mens have to be worked over, that all processes 

 should be simple and expeditious as well as 

 accurate. Life is not long enough to do it as 

 accurately as one might wish, but I have used 

 the following rough-and-ready method with some 

 success. Each band has three grades of strength : 

 full strength represented by, say, 3 ; half strength, 3 ; 

 quarter strength by f . Strength above full strength 



