72 Miscellaneous. 



a passage through earlier stages of the development of instinct, such 

 as has long heen known in the development of organs. 



It has often been asserted that the geometrical spiders do not 

 repair old webs. This, however, is true only in a limited sense. 

 The outer framework and some of the radii which have already 

 become nearly free from transverse threads are probably always 

 used again by Z'dJa oc-notata and others. The rest is gathered up, 

 worked into a ball with the mouth, and thrown away. If the 

 spider removes a lifeless object from the web, and damages the latter 

 in so doing, it certainly sometimes reproduces the destroyed portion 

 of the framework, the radii, and the central shelter. If we interrupt 

 a spider in the formation of its web, by tearing away a portion of it 

 with the corresponding part of the outer framework, all will be 

 completed up to the part that has remained uninjured. In this 

 case the completion of the framework is especially interesting, as 

 this unaccustomed work is not usually successfully performed at 

 once. Here we see very distinctly how reflection comes into play. 

 I was still better able to ascertain reflection, or, what is the same 

 thing, actual inference, in the case of Attus areuatus, BL, when I 

 offered it flies touched with oil of turpentine. Sometimes the spider 

 despised the species of fly employed (ffomalomyia canicularis, L.), 

 whilst it attacked other insects (e. g. Chironomus tendens, Fab.) just 

 as before. This spider also draws similar conclusions in those cases 

 in which it cannot overcome insects in consequence of their chitinous 

 armour being too hard. These it usually attacks only once, and is 

 then for a long time forewarned. Dangerous insects, however, such 

 as small bees, it avoids, without having seen their sting. Here there- 

 fore we have an instinctive dread. Bee-like flies are equally dreaded. 



I have also attempted to give a new explanation of the secondary 

 sexual differences of many spiders, which are to be ascribed to 

 changes by means of sexual selection. — Zool. Anz. no. 180, p. 591. 



On the Classificatory Position of Hemiaster elongatus. 



To the Editors of the Annals and Magazine of Natural History. 



Gentlemen, — You were good enough to admit a reply on the part of 

 Mr. Percy Sladen and myself to a criticism of Prof. Sven Loven, upon 

 the classificatory position of Hemiaster elonyatus, nobis, in your numher 

 for October last. I have received, in consequence, a very cordial 

 reply from the Professor, in which he acknowledges that the form 

 is not a species of Palceostoma, and points out how these latter 

 forms of Hemiaster depart from the Mesozoic types of Desor, Wright, 

 and Cotteau, in the extension of the madreporite and in the dimi- 

 nution in the number of the ovarial pores. He suggests that we 

 should place our species in a new genus. The consideration of this 

 proposed splitting up of the genus Hemiaster we must defer for a 

 while, for it is a matter that concerns M. de Loriol also ; and, more- 

 over, we can hardly determine the propriety of the step until we 

 have completed our description of the Echinoidea of the Tertiaries 

 of iSiud. 



Tours &c, 



Dec. 1, 1884. P. Martin Duncan, 



