556 H. M. Bernard — Developing Trilohites. 



bottle. This bappy device brought the whole available force of the 

 air upon the sand, and its cutting action vv^as much increased. Mr. 

 Last has kindly iurnished me w^ith a couple of sketches explaining 

 the apparatus as it w^as used. In the section, a represents the level 

 of the sand, nearly run out ; h, the feed-pipe ; c, the tube carrying 

 the air above the sand ; at <Z is a wire ring which, by contracting 

 the air passage between the two vertical pipes, secures a slight 

 excess of pressure within the bottle. Without this excess of pressure 

 the sand was not fed into the jet in sufficient quantity ; e is the 

 nozzle, formed by a piece of fine bore india-rubber tubing; /is the 

 wire clip to keep the cork from being blown out. 



A considerably larger bottle might have been used with advantage. 

 The one figured only lasted about five minutes ; although it was 

 easily filled, yet, when there was much matrix to be cut away, it 

 would have been more convenient had it lasted 15 to 20 minutes. 



The fossil was embedded in paraffin, for convenience of manipula- 

 tion and for the ready protection of exposed parts, the under-surface 

 alone being exposed. It was then fixed in position by screw clips 

 in front of the nozzle of the sand-blast, and the whole covered with 

 a large glass shade, to prevent the dispersal of the sand. The shade 

 stood on feet so as to allow the air to escape. Parts which at any 

 moment required protection were covered with the paraffin by means 

 of a hot knife or needle. Various other methods of protection were 

 tried, such as collodion, india-rubber solution (this applied by a 

 fine brush), strips of gum-paper, small cardboard screens, but none 

 of them helped much. 



Although I devoted many hours to the process, I was too busy 

 with other things to be able in any way to exhaust the method ; on 

 the contrary, I did but make a beginning, and cannot, therefore, 

 pretend to give a final judgment on its merits. At the same time, 

 my experience of it, as far as it went, may not be without interest. 



The differential action is not so great as I expected it would be, 

 still it undoubtedly exists. Many times structures appeared which 

 looked very like fragments of limbs. These I did not care to try 

 to save. I was looking for a series of segmentally arranged pro- 

 cesses which ought to appear in more or less regular rows on each 

 side, hence single fragments were of little use to me. The interesting 

 question unfortunately remains to be solved : " Had such rows of 

 limbs come into view, would the sand-blast have differentiated them, 

 or would it have cut them away together with the matrix?" My 

 experience, as far as it goes, makes me believe that this would 

 depend entirely upon the hardness of the matrix. In one partly 

 rolled specimen (such specimens, inasmuch as they must have been 

 complete animals when they rolled up, are the most hopeful to 

 experiment with), I found with a pocket lens, a regular row of 

 dots close under the lateral edge of the head-shield. I worked away 

 the matrix, commencing near the middle line and working slowly 

 outwards towards these points so as if possible to reveal the limbs 

 along their lengths. As I suspected, the points were the tips of 

 structures which were almost certainly limbs, but the matrix was 



