210 DR OTTO THILO. 



nest where the careful father watches over them, and only 

 too often the stickleback succumbs to the attacks of his many 

 wives. 



In order to meet these many-sided attacks the spines 

 of the stickleback are provided with click joints which 

 admit of much elasticity in their use. As quick as lightning 

 a stickleback can force up his spines if he is irritated, and 

 can hold them up for hours at a time without any strain on 

 the muscles. One is easily convinced of it, when one tries 

 to lay down the upright spines of a dead stickleback. If 

 you press against the point of the spine it is impossible 

 to lay it ; on the other hand, with the point of a pin one can 

 lay the spine, if pressed on the proper spot, on the front 

 part of the joint end. This surprising fact we can under- 

 stand when we have thoroughly examined the joint of the 

 spine and its click mechanism as compared with those of other 

 fish. The click mechanisms are very varied, and some of them 

 are difficult to understand. I have seen in some large 

 museums small fish in very large bottles because they had 

 not understood how to lay the fin rays to enable them to 

 be put into bottles corresponding to their size. 



The Click Mechanism of the Stickleback — The spine of 

 the stickleback shows on the joint end under a split (fig. 12) 

 an eloncration of the bone, which is formed like two bent 

 swords. These elongations rest in sheaths resembling bent 

 swords when pushed into their sheaths. A rough sketch of 

 the stickeback spines, taken from the side, is given in fig. 6. 

 The result of these conditions prevents the spine from being 

 pushed into a sheath by force from K to S'. These move- 

 ments towards S' can only be worked by a force K', which 

 describes a circle A' B K'. The lock acts only when the spine 

 lies so that its prop is at the point A and B. If you change 

 the prop A to A', one can without hindrance move S to S', 

 then with the diameter A' B one can make a circle which lies 

 tangential to the circle A' B K'. Should, on the other hand, 

 the prop be at A, the circle A' B K is crossed by a circle that 

 is described with A B. 



The correctness of these explanations can easily be proved 

 by a model in wood or iron. 



