378 Mr MUNRO, on A METRICAL 



modulated the speech of man, as to place on every word one acute tone, and not more than 

 one, and that one not beyond the third syllable from the end.' In modern times many have 

 found it impossible to conceive what he thinks it impossible not to conceive. Such crea- 

 tures of habit are we. As to the limits within which the accent might range, the two lan- 

 guages are agreed ; as to the place it might have within these limits they differ greatly. 

 In words of more than one syllable, with few and peculiar exceptions, the Latin accent was 

 never on the last syllable. In this respect it departed widely from the other Greek dialects, 

 but agreed curiously with the Aeolic, with branches of which dialect in Italy the Latins were 

 so long in contact. But in another and even more important point the Latin was in direct 

 opposition to the Aeolic, as well as all other Greek dialects. In Greek the length of the last 

 syllable limited the range of the accent ; the length of the penultimate made no difference 

 whatsoever. In Latin polysyllables the length of the last syllable was quite unimportant ; the 

 length of the penultimate absolutely determined the place of the accent. If it is long, the 

 accent must be on it, if short, it cannot be on it. To give a few examples. 



Monosyllables in which the vowel was long by nature, were circumflexed ; as sol, ros, 

 mos,pons, mons, res, os (oris), est ('eats'). Those in which the vowel was by nature short, 

 were oxyton ; as mil, cor, vir, mors, nuw, os (ossis), est (' is'). 



Dissyllables, the penultimate of which was short or only long by position, were paroxy- 

 ton ; as arma, virum, vinit, deos, esse ('to be'), essent, lectus ('bed'). Those in which the 

 penultimate was long by nature, if the final syllable was also long either by nature or position, 

 took the same accent; as oris, fato, Romae, celant. But if the last syllable was short, the 

 penultimate was circumflexed; primus, venit, iram, musa, lectus (particip.). 



Polysyllables, if the penultimate was long either by nature or position, had the accent 

 on that syllable ; and whether that accent was a circumflex or acute, was determined by the 

 same rule as in dissyllables : regina, adire, pietate took the circumflex ; inferret, Albani, 

 labores the acute. If the penultimate was short, all polysyllables, whatever the quantity of 

 the antepenultimate or of the last syllable, were proparoxyton ; as Italiam, profugus, litora, 

 caelestibus, asperrimi. 



The following are exceptions to these general laws. The enclitics que, ve, ce, ne attract 

 the accent to the syllable immediately preceding, whether long or short : armdque, as well as 

 armisque ; illdve, istoce, sicine. When ce and ne suffer apocope, the accent is then on the 

 last syllable : illUc, adhuc, istoc, audin, viden, tanton, crudelin. In a few other cases too 

 the accent is on the last syllable, as in nostras ('of our country'), vestras, cujas, Antias. 



The atonies as they are called, that is words so closely joined with another that they 

 become as it were a part of it and lose their own accent, are much more numerous in Latin 

 than in Greek; comprising all the prepositions, many conjunctions, and the relative, not the 

 interrogative, qui, quae, quod. Particles too are often joined enclitically to the word pre- 

 ceding them. Quintilian quotes from the first line of the Aeneid quiprimus aboris, where both 

 qui and ab are atonic, that is to say really form but one word with primus and oris respec- 

 tively. An ancient Latin seems to have been able by the sense alone to distinguish in justo 

 from injusto; or praeter missa from praetermissa, even dissyllabic prepositions being atonic. 

 Of circum litora Quintilian says that some grammarians taught that circum, like the Greek 



