TRANSACTIONS OF SECTION D. 899 



But this is not all. During the forty or fifty years which have gone by since 

 the breakwater was completed, the towns on the shores have largely increased 

 their population ; the great dockyard at Devonport has increased in size and in 

 activity ; and the ships which visit the Sound are larger and more numerous than 

 they were. Now the sewage and other refuse from these great and growing towns 

 and dockyards, and fi-om all these ships, is thrown into the Sound ; so that while 

 it is more difficult than it used to be for fine silt to be washed out of the Sound, 

 the quantity thrown into it is much greater than it was, and is becoming greater 

 every day. 



It is well known that these changes in the physical conditions of the Sound 

 have been accompanied by the disappearance of animals which used to live in it, 

 but which are now found only outside the area affected by the breakwater. 



These considerations induced me to try the experiment of keeping crabs in 

 water containing fine mud in suspension, in order to see whether a selective 

 destruction occurred under these circumstances or not. For this purpose, crabs 

 were collected and placed in a large vessel of sea-water, in which a considerable 

 quantity of very fine china clay was suspended. The clay was prevented from 

 settling by a slowly moving automatic agitator ; and the crabs were kept under 

 these conditions for various periods of time. At the end of each experiment the 

 dead were separated from the living, and both were measured. 



Fig. 5. 



s n 



30 



20 



10 



30 



20 



10 



_50 -40 -30 -20 -10 +10 +20 +30 +40 



Diagram showing the effect of china clay upon 218 male crabs. The upper curve 

 shows the distribution of frontal breadths in all these crabs ; the dotted curve the 

 distribution of frontal breadths in the survivors. The dotted line s shows the 

 mean of the survivors ; the line D the mean of the dead. 



In every case in which this experiment was performed with china clay as fine 

 as that brought down by the rivers, or nearly so, the crabs which died were on the 

 whole distinctly broader than the crabs which lived through the experiment, so 

 that a crab's chance of survival could be measured by its frontal breadth. 



"When the experiment was performed with coarser clay than this, the death- 

 rate was smaller, and was not selective. 



I will rapidly show you the results of one or two experiments. The diagram 

 (fig. 5) shows the distribution of frontal breadths, about the average proper to 

 their length, in 248 male crabs treated in one experiment. Of these crabs, 154 

 died during the experiment, and 94 survived. The distribution of frontal breadths 

 in the survivors is shown by the lower curve in the diagram, and you see that the 

 mean of the survivors is clearly below the mean of the original series, the mean of 

 the dead being above the original mean. 



Two other cases, which are only examples of a series in my possession, show 

 precisely the same thing.^ 



These experiments seemed to me to show that very finely divided china clay 



' It is impossible in this place to give a full account of the experiments referred 

 to, and a multiplication of mere small-scale diagrams seems useless, so that only one 

 of those exhibited when the address was delivered is here reproduced. 



3 H 2 



