Introduction. 



The purpose of this treatise is to present the results of an in- 

 vestigation into the nature of the rhythm that characterized the 

 prose of Ammianus MarceUinus, the great historian of the fourth 

 century. Inasmuch as the subject is by no means an entirely new 

 one, it is fitting that we should begin by passing in brief review 

 the facts already established. 



The fundamental features of the rhythm employed by Ammianus 

 were pointed out some years ago by Wilhelm Meyer of Speyer in 

 his review of Havet's book, La piose uu'trique de Symmaque} Meyer 

 discovered that in common with a large number of authors, both 

 Latin and Greek, from the fourth century on, Ammianus makes use 

 of accentual clausulae, — in other words, that he secures a rhythmical 

 effect in his prose by contriving that each clause shall end in one 

 or another of a limited number of accentual cadences. As a general 

 definition of the character of these cadences Meyer formulated the 

 following rule : 



Both in the Greek and in the Latin accentual clausulae the rule 

 holds good that either two or four (rarely three) unaccented syllables 

 separate the accented syllables of the last two words in the clause. 



Now this is a sweeping generalization, and as such it has the 

 faults that are consequent upon its merits. For in the first place, 

 although it holds good in the case of many (probably the majority) 

 of the Latin writers, there are on the one hand some who entirely 

 avoid the use of clausulae in which three syllables separate the 

 accents, and on the other hand some, like Ennodius, who employ 

 these cadences even more extensively than those in which there 

 are four intervening syllables. And in the second place, granted 

 its general truth, it remains an inadequate definition of the Latin 

 clausulae, for it leaves out of consideration two important factors, 

 to wit, the question of the number of the syllables which follow the 

 last accent, and the matter of word-division.^ 



» Gott. gel. Am. 1893, 1-27 = Ges. Abh. ziir Mittellat. Rytlumk II, 236-286. 

 References to Meyer will hereafter be made to the reprint. 



* It is not in a spirit of carping criticism that I bring up these points, 

 but simply as a protest against a growing tendency to view this state- 

 ment as a law of Latin prose-rhythm, whereas it is nothing more than 

 a convenient expression of the extent to which Latin usage, taken by 

 and large, agrees with Greek usage. 



