THE ISNABE 01 THE GIAST WOOD SPIDER. 
915 
segment, and let us now look to the diagram (Fig. 2) and witness what has 
occurred. 
Fig. 2. — Loss of paraiielism resulting fiom division of one turn 
of viscid spiral in one segment. 
(a) Viscid spiral divided in segment X. 
(J) Result of spider’s work. Loss of parallelism in both 
• X and Y. 
Arrow mark the direction of spider s circle. 
Turn first to diagram a. Let w, x, y, z be four adjoining segments. Let x 
be the experimental segment, and let three turns 1, 2, 3 of the viscid spiral be 
complete. I divide the innermost turn in segment x. Now turn to diagram />. 
The spider circles on laying down the fourth turn of its spiral all round the snare. 
At length it reaches the radius between z and y. Here all is well. It stret- 
ches out its hind leg ; it touches the radius between the segments exactly at 
the jxrint o, that is the point where the radius meets the last turn of the viscid 
spiral. This point of measurement is intact. Therefore the spider makes an 
accurate judgment, and draws her line parallel across segment z. She passes 
on to the radius between x and y. She again stretches out to hook the radius 
.at the point wLere it meets line 3. But line 3 is gone, therefore she must touch 
the radius where it meets line 2. Her measurement is consequently incorrect. 
Her attachment is made in the wrong place, farther out on the radius, and 
the line drawn across segment y is out of parallel. The spider passes on. She 
reaches the radius between w and x. She stretches forward ; here she finds 
the line correct, and she anchors her filament in the right place. But the other 
end of that filament has found a wrong attachment, therefore the line across 
segment x is also out of parallel. To sum up ; the result of the experiment is 
this. One turn of the viscid spiral is removed from one segment. The point 
where it touches the radius is as a consequence destroyed. But, since this is 
the measuring point for the spider, therefore the next turn of her viscid spiral 
is incorrectly laid. Its parallelism is lost in both the experimental and the 
preceding segment. In the one the lines converge, in the other they diverge. 
The essential fact is that the point of measurement being removed then the 
parallelism is lost. Thus the clear conclusion is that each attachment of the 
viscid spiral is measured off from that point on the radius where it is touched 
by the preceding turn. 
So much is confirmed ; but we have not yet clinched the fact that the hind 
leg is the measuring rule employed. We must make certain of this observation, 
since it is directly opposite to the method of the Aranevs which measures by the 
aid of her fore leg. I satisfied myself of this by first amputating the legs 
and then witnessing the result. Let us attempt a similar experiment on the 
Eephila. But in her case it is the hind legs which we must first cut away. It 
