232 SUPPLEMENT TO 



sucli as that of Jtj/pus, and the funnel nest of Cyrtait' 

 chefims elongatus, neither of which is furnished with 

 a door. 



Among the true trap-door nests, those of the cork 

 type stand in a measure alone, being distinguished 

 from all the others by their solid surface doors, com- 

 posed of many layers of silk and earth ; and we do 

 not at present know of any intermediate forms linking 

 the cork and wafer types together. But among the 

 various nests which represent the wafer type the case 

 is different, for here the types naturally fall into a 

 progressive series, such as that represented in the 

 diagrams (PI. XIY., p. 193). 



If we try to picture to ourselves the stages through 

 which the most complicated wafer nest — namel}^ that 

 of the double-door, branched, cavity type (Diagram G 1) 

 may have passed in the course of its development 

 from a simpler ancestral form, we should a priori 

 expect to find precisely such structures as the Hyeres 

 double-door branched nest (Diagram F), and the sinyle- 

 door branched nest (Diagram D) forming successive 

 halting-places in the advance from the primitive 

 single-door, unbranched nest (Diagram C). 



The double-door unbranched type may in like manner 

 find its prototype in the same original single -door un- 

 branched nest (C), which we may look upon as the 

 parent idea, from which all these structures have been 

 derived. 



Bearing this in mind, and remembering that kin- 

 ship between living creatures is not only revealed to 

 us by likeness in structure and colour, but also by 

 similarity in habits and instincts, it becomes of interest 

 to trace any resemblance that may exist between 



