18 Mr. II. M. Bernard on tlie 



than 13/. And again variations of curve are seen in the 

 figured section of Chonophyllum perfoUatum * with proto- 

 theca 13 h. 



Thus at the very outset we find ourselves face to face with 

 the crux of all systematic work : What is the taxonomic 

 value of these slopes and curves of the rim in any individual 

 case? We know from Mr. Pace's observations t that great 

 variability in the openness and flatness of the calicle can be 

 correlated with the degree of muddiness of the water. The 

 sediment runs more easily off a coral with a flattened open 

 theca than from one with a cup-shaped theca. Then, again, 

 we are justified in assuming that these forms were developed 

 in each case by slow modifications of an originally deep pro- 

 totheca (age, therefore, may have something to do with the 

 form) ; and, lastly, we can imagine many different accidents 

 which would tilt or depress such rims. 



Nevertheless we have a structure of fundamental importance 

 in the coral skeleton, and the form-variations of this structure 

 may justly claim to take the first taxonomic rank. But how 

 are w T e to distinguish those of importance from those which 

 are accidental in individual cases ? The matter is further 

 complicated in the case of these ancient fossils, because the 

 transition-forms are preserved equally with those which have 

 passed over finally to some well-defined type. It seems 

 fairly clear that classification of such forms must be attempted 

 on wholly different lines from that still in vogue. Before 

 any form receives a name we should satisfy ourselves by a close 

 study of series that it embodies some new principle of struc- 

 ture. Three or four such distinct principles can be gathered 

 from the forms of the prototheca given in PI. I. diagram ISa-g. 

 In a the rim continues to show no sharp bend downwards, and 

 is distinct from that in which the rim tends to bend out so as 

 to form an open dish either as 13 c or 13/; and both these 

 differ from the sharper curve of the edge all round (13#). 

 Fig. 13 h, in which the edge bends rapidly over and then either 

 hangs straight down or shows a tendency to curve up again, 

 seems to me to be very easily distinguishable from 13/, for, 

 even though the two might possibly pass into one another, a 

 smooth curve and a sharp bend are very distinct. 



I propose now to leave all but one of these early variations 

 of the prototheca, hoping that I have said enough to claim 

 greater consideration for them in all future work on Palaeozoic 



* Brit. Foss. Corals, pi. 50. fig. 5. The section perhaps does not run 



true. 



+ Ann. & Map-. Nat. Hist, ser, 7, vol. vii. (1901) p. 385. 



