H. M, Bernard — Developing Trilohites. 557 



so hard, that no differential action was possible, and further, the 

 faint ridge of darker material which represented the limb first 

 revealed was too thin to protect with parafiin. 



On the other hand, if the matrix is not very hard, the sand-blast 

 would almost certainly differentiate tine structures, provided, of 

 course, they were not exposed too long to the action of the sand. 

 My final plan, therefore, was to work away till I discovered by 

 means of the sand-blast the existence of a serial row of limbs in 

 a specimen, and then to blast away the matrix on all sides of these 

 limbs, even through the dorsal surface, always stopping as soon as 

 they showed as ridges, and then to work in some other way to 

 clear the limbs further. Unfortunately, I never found such a 

 specimen, and believe that it would be very difficult to find. The 

 rolled-up or partly rolled-up specimens get filled with a substance 

 too hard for the satid-blast to yield any result, while the limbs of the 

 open specimens were probably flattened down against the dorsal 

 carapace into a tangled mass. 1 did not, however, give up hope 

 that, if I could work long enough at it and had sufficient material, 

 the sand-blast would reveal something worth finding, if not a complete 

 series of limbs. As above stated, many m^'sterious fragments and 

 membranes appeared which kept me for a few days or hours iu 

 a state of tension. For instance, a fragment of the ventral mem- 

 brane was very clearly shown in one specimen (figured in the paper 

 above quoted). 



The absence of nearly all tangible result from my own researches 

 need deter no one from repeating the attempt. It must be re- 

 membered that a very large portion of the time I devoted to it 

 was given to getting an apparatus which would work, and then 

 when this was overcome by the mechanical skill and kindly assist- 

 ance of my friend Mr. Last, other and more pressing duties called 

 me away, so that on the whole I did not apply the method to more 

 than six specimens in all. From every one of them I think 

 I learned something, and I firmly believe that if anyone with more 

 leisure and sufficient material would repeat the attempt, he would 

 sooner or later certainly be rewarded. We have not by any means 

 exhausted all that could be learned from a study of the inner surface 

 of the dorsal shell, which the sand-blast is well able to clear out. In 

 these cases, it is well to take care that the carapace is firmly bound 

 together by strips of gum-paper, or by embedding in paraffin, 

 otherwise it is very likely to fall to pieces as soon as the matrix is 

 removed. 



It may be said, in conclusion, that when the matrix is soft as 

 compared with the fossil — and this would apply to all fossils — the 

 sand-blast cleans the objects very beautifully, and for this alone, 

 as an accessory method of investigation, it might be of great use 

 to palgeontologists. It is only with great regret that I have felt 

 myself compelled, on account of other duties, to drop this method 

 of investigation, and I hope that some one with sufficient leisure 

 and opportunity will be tempted to carry it on. 



