Since Niebuhr the opinion has generally prevailed that the curia was composed of gentes. A passage which at first glance seems to have a bearing on the question is Dion. Hal. ii. 7. 4: “Romulus divided the curiae into decades, each commanded by a leader, who in the language of the country is called decurion.”[75] The word decurion proves, however, that in speaking of decades Dionysius is thinking of the military divisions called decuriae, each commanded by a decurion. In historical times the troop of cavalry—turma—was divided into three decuriae of ten each, as the word itself indicates. There were accordingly three decurions to the turma, and ten turmae ordinarily went with the legion.[76] From Varro[77] we learn that the three primitive tribes furnished turmae and decuriae of cavalry, the decuriae commanded by decurions. Dionysius accordingly refers to military companies—either to the well known decuriae of cavalry or to corresponding companies of footmen which probably existed before the adoption of the phalanx.[78] Had he meant gentes, he would have used the corresponding Greek word γένη. Niebuhr[79] inferred from this passage that each curia was divided into ten gentes, making three hundred gentes for the entire state; but a careful interpretation shows that no reference to the gentes is intended. We cannot infer therefore from this citation that the curia was divided into gentes. The other passage relative to the question is Gellius xv. 27. 4,[80] in which Laelius Felix states that the voting in the comitia curiata was by genera hominum in contrast with the census et aetas of the centuriate assembly and with the regiones et loca of the comitia tributa. Niebuhr identifies genera with gentes.[81] It is clear, however, that in this passage Laelius is not concretely defining the voting units of the various assemblies, but is stating in a general way the principles underlying their organization into voting units. In the comitia centuriata the principle is wealth and age; census et aetas is not to be identified with centuria or with any other group of individuals in this assembly. In like manner regiones et loca expresses the principle of organization of the tribal assembly; or if used concretely, it must designate the tribes themselves, and not subdivisions of the tribes, for none existed. Correspondingly genera hominum signifies that the principle of organization of the curiate assembly is hereditary connection; but so far as the expression is applied concretely, it must denote the curiae themselves not subdivisions of these associations. The curia, a religious, social, and political group based on birth, might well be called genus hominum in contrast with the local tribe and with the century, composed artificially of men of similar wealth and age. It is well known, too, that voting within the curiae was not by gentes but by heads.[82] As no other passage from the sources, besides these two, has