438 H. B. Dewing, 
The case is irregular, and no good cursus type is possible with or 
without elision; but the punctuation may be questioned. 
(A. nai nj wiv eiyeto, 4 Dé Ewedler, 2805 A. 
Here ciyeto, 4 dé Zueddev scans Correctly: + _/_/_/+ __3 but eiyer, 
7 0" epediey is equally good: jj. 
5. atoatnyay TE oxhos, 2808 B. 
Here the text is corrupt. 
The following three cases show elision as printed (in each case 
of a preposition) : 
6. Te weylota nae’ avdounas, 2801 A. 
7. wo eixdc, én’? adbt@- 2816 A. 
8. adhe ta xa¥ Huds. 2817 C. 
In all these the cursus requires that the elided forms be retained. 
To sum up: Procopius of Gaza very carefully avoids hiatus. 
Only one clear case appears (2) in which a hiatus avoidable by 
elision must be retained to keep the cursus, against four cases 
(1, 6, 7, 8) where elision must certainly be admitted, once against 
the manuscripts (76 xzju’ édoyitov). Procopius of Gaza is plainly more 
careful than Zosimus in admission of hiatus, but one clear case 
forbids the conclusion that elision must be introduced to avoid 
hiatus wherever possible. Both writers are certainly inconsistent in 
regard to elision in their clausule. 
In the Political Verses the neglect of elision is a matter which 
varies with different writers. In Constantinus Manasses, vv. 2500—2700 
in the Corpus, there occurs no case of hiatus, while there are oc- 
casional cases of elision and crasis. One case of hiatus was found 
in another passage (v. 2306 dvo éry), but it is clear that this writer 
does not admit hiatus of any sort except in very rare cases, and 
he consistently elides when hiatus can be avoided in this way. 
In John Tzetzes, Historiarum Variarum Chiliades, ed. Kiessling, 
I, 1-204, seventeen cases of hiatus occur; of these five are justified 
by their position in the ceesura of the verse. One is a quotation of 
a familiar phrase: te ovov 130. 
Four are cases of weak hiatus: 
zal ovowud 67. 
Koavaod tacoyov Ie}: 
zat “Adnrre, 173. 
tovtov ‘Advattns, 155. 
Seven cases of plain hiatus are found: 
