Craniolof/ical Noten 



341 



Let US iiow sec oxactlv in wliiit l'nilossor v. Türük'.s critiuism coiisist.s. It contaiii.s tliren 

 separate cliai'ges agaiust the aritliiuetic iiicaii : 



(i) It does not give tlio rul(-, somctliin>,' cliaracteristic of evcry memlicr (if a group tenncii 

 hy craniologists tlie typo. 



Obvioii.sly and i'loarly it canni.t, and il' aiiy uraniologi^t thinks it dües, lie »hould bcgin hi.s 

 Studios ah initio with a rcading uf Qui'tclet'.s woi-ks. 

 (ii) It is not idcntical with the iiiock'. 

 (üi) It is not idcntical witli the medioti. 

 These defects (i), (ii), aiul (iii) niake tlic nican uf no Service at all in craniological discussions. 



Nüw Ict US considcr (ii) and liii) in the light ol' Professor V. Türiik's own data tnuited hy an 

 adequate Statistical theory. Table I. gives Ins data for greatest forehead breadth (S. 509) and 

 for «Toatest skull brcadth (S. 509). I sclected these two seriös out of the four givcn because they 

 lookcd in Professor v. Tiirük's diagrams the " skowest,'' and therefore, if there was a sensible 

 di.stinction botwecn nioan, modo and median, I thought it would certainly be evidenced in these. 



TABLE I. 



Now these seriös are so nearly symmetrical that it seenied sufficient to graduate them with 

 a curve of the type : y = v/„n+-) ('-yr, whore p = ya : seo Plill. Tnins. A., Vol. 186, p. 367. 



