148 M^'aUer Peterseu, 



7,ixavrj, BeppiT/.c-r : J^spp'-r, [jxipa/iT/.o? : [jxipaxiov, o?/i'7XY, : ot/i:(, /e- 

 Ai'Txov : /sXetov, zrfiy\.fjv.oz : ~7tyoc,^ [iac-dioy.o; : jiaT'.As'J?, aixcpopiTxo? : 

 a|j.'^op$tj?. The <t stems followed the analogy of the vowel stems, 

 since the loss of the t in the majority of forms and the subse- 

 quent contraction of vowels caused their suffix to be no longer 

 felt as the stem-final : therefore : o-cpEliTxov : o-cpsT^a?, oVwsXtTxov : 

 'jy.tkoc, [j-eXtT-xov : [xilo^. -w- stems on the other hand seem to 

 have retained their w and added -toxo- of two full syllables : 

 'Hptoi-rxoc ; rjpo);. In several instances consonant suffixes were also 

 dropped : xuXi(7XYi ^ (by haplology for ^xLtXixiTXY] ?) : x'jXt'l, and regularly 

 -tB-, e. g. c-avi«7>tY] : aavi;, xotcitxo? : xotcic. We may assume that 

 words like >«£xavt'7X7] : Xexavr,, but with };£xavi; existing alongside, 

 sometimes caused the word in -iaxo- to be referred to the primitive 

 in -t;, and the others followed by analogy.^ The reason for the prefer- 

 ence of the short form lay in the avoidance of awkward forms like *xo- 

 ziBioVwOC. More obscure is Tpi[j.i(yxo? : Tpi[j.t,T&?, though possibly the word 

 in -iTO- could have dropped the to by analogy to words in -iB-. For 

 Ku(i£pvi<7xoc : xul^spvTjTYi? see note to § 103 A. Finally there are 

 several cases of the dropping of an n suffix before -tcrxo-. So cer- 

 tainly the Plautine Olympiscus and Lampadiscus, which act as 

 living ' diminutives ' to Olympic and Lampadio (§ 102.) Prob- 

 ably also Mevicrxo? : Msvo>v. The cause for this again may have 

 been the existence of a collateral form with an n suffix alongside 

 of the real primitive without it, e. g. the existence of Ilappivcov 

 beside Ilapjjivto?, so that nap[j.svi<Txo? could be referred to the 

 former instead of the latter. Similarly IToOaaxo? : floO-o? could be 

 referred to lloOuov, and cause by analogy Aa[j.7:aBi(7xo? : AajxTraBiwv, 

 etc. Yet in view of the fact that all of these words are proper 

 names it seems more probable that our suffix has taken upon it- 

 self the function of forming " Kosenamen " or shortened forms of 

 proper names. Cf. Brugmann, Gr. 2. i-. 119. Certain of this we 

 are for two compound names the last component of which is al- 

 most entirely suppressed before the -loxo- : IlapixiTXOi; : Hap^XEvwy 

 and llav-c-a}i<7xo$ : IIavTa>v£o>v. See § 105. 



^ On the other hand (xviaxos : fxvg, &viaxos : 9vog, where the v is part of the 

 root. Influence of other words in -iaxo- prevented the contraction of vi ">■ v. 

 &%'axr} in Hesychius, however. 



" Janson, De Graeci Sermonis Dem. in iaxos 6, assumes *xv'/.r} as prim- 

 itive, though that it hardly necessary. 



^ These words were certainly not syncopated, as is assumed by Janson, 

 op. cit. 4, for ;fArt»'i'«Tx<o»'. 



