Before the censorship of Appius Claudius Caecus, 312, military service within the census ratings was based on the possession of land, and the gradations of equipment, while they lasted, must therefore have been determined by the size of the estate reckoned in iugera.[500] Huschke[501] rightly inferred that the number of iugera marking the lower limit of each division must have been proportioned to the later money ratings, and assumed accordingly 20, 15, 10, 5, 2½ or 2 iugera as the respective minimal holdings of the five divisions. Although absolute certainty is unattainable, most scholars accept his conclusions as probable.[502] Before the change was made in the appraisements from amount of land to money, the census gradations ceased to serve a military purpose. In the further discussion of these groups reference is therefore solely to their political character, especially as expressed in the organization of the comitia centuriata. Till the time of Marius, however, the soldiers were ordinarily recruited from the classes—that is, from the citizens who possessed at least the qualification of the lowest group.[503] The money ratings of 312 are not recorded; we know those only of the time following 269. The ratings of the earlier date must have been in the nominally libral asses then current. For a long time, probably down to 312, the _as_ remained at eleven to nine ounces in weight, then sank rapidly to four, three, and two ounces, reaching the last-mentioned weight in or shortly before 269. In this year or the following was legally adopted the lighter _as_, weighing two ounces, or a sixth of a pound, and hence termed sextantarian, and the heavier asses still in circulation were henceforth reckoned as sesterces, which now became the unit of value.[504] Two and a half sextantarian asses made a sesterce, and four sesterces made a denarius.[505] The _as_ continued to be copper, whereas the sesterce and the denarius were silver. In consequence of the use of the sextantarian _as_ the ratings must have been elevated to correspond with the decline of the standard; and the result of this change is the well-known series, 100,000, 75,000, 50,000, 25,000, 11,000.[506] There can be no doubt that under the standard used in 312 the ratings were lower than those given. It is incredible that the classis should ever have been appraised so high as 100,000 asses of ten-ounce weight or even of the value of sesterces (5 oz.).[507] But the ratings of 312 have not been definitely ascertained. Assuming but one elevation between the two dates and in the proportion of 4:10:: sextantarian _as_: heavy _as_ or sesterce, Mommsen[508] concludes that the appraisements of 312 were 40,000, 30,000, 20,000, 10,000, and 4400 asses respectively for the five classes. The adjustment however may have been gradual, as was the decline of the standard, and the former need not have corresponded exactly with the latter. But in so far as the Romans