W. K Macdonell 
307 
it is in some degree due to the operation of the ordinance of the Emperor Marcus 
Aurelius, who was the first to introduce an official registration of births in Rome 
and the Provinces for the special purpose of establishing the status and age of the 
people when required*. 
But, as I have said, in the great majority of the inscriptions the age is given 
in years only, and in many cases it was evidently not known accurately, probably 
no record of birth having been kept, or the record having been lost or destroyed. 
An examination of Table I will make this clear. It will be seen that after the 
15th year the frequencies are heaped up on ages expressed by multiples of 5; 
e.g. at 25 years we find 129 deaths of males, but at 24 and 26 only 57 and 
34 deaths respectively. Soldiers' ages are treated in the same way. Perhaps 
the explanation is that in many of these cases the age was not known exactly 
in years, but that the number of lustra of five years which the deceased had lived 
would be known fairly correctly, and would thus furnish the basis on which the 
number of years was calculated. Of course we cannot say how such a method of 
computation, if it existed, would be worked out in practice ; if for instance it 
was known that no long period had elapsed since a person had completed his 
6th lustrum, his age might conceivably be recorded as 30, if a considerable period 
had elapsed, as 35. Whatever the real explanation may be, it is evident that 
the practice causes a difficulty which will be obvious from an inspection of the 
diagrams on Figs. 1 — 3 : the polygons showing the expectation of life in the 
ancient world present a jagged appearance in marked contrast to the smooth, 
continuously descending curves which represent the data for modern times. 
But these data are subjected to smoothing before publication. The difficulty 
might have been got over, as Mr Palin Elderton has suggested to me, by adding 
together the deaths recorded at ages 18, 19, 20, 21 and 22, calling these 20 
(and similarly for the other groups) and then redistributing them. The same 
result, however, can be attained fairly accurately and with a saving of labour, 
by supposing the jagged polygons replaced by continuous curves, and such curves 
can be readily placed on the diagrams. 
Date of the Inscriptions. Few of the inscriptions which record age at death 
give us a direct clue to the dates at which they were erected. Some that were 
found in the "Monumenta" of families and households of historical personages 
(Livia, Nero, Drusus, Marcella, etc.) can be assigned to the early Empire, and 
a few, mainly Christian of the 4th to 6th centuries, give the names of the 
contemporary consuls, and can thus be dated exactly. In one case, No. 9919, 
the name of the Emperor, Mauricius (a.d. 582 — 602), and the year of his reign 
are given. It is possible that scholars can date the inscriptions from the form of 
the letters, the spelling, etc., but as I am not aware of this having been done, 
I must leave the point unsettled and would suggest with some hesitation that 
the great mass of the inscriptions belongs to the first three or four centuries 
of our era. 
* Marqnardt, Das Privatlebcn der Rimer 2 , p. 80. 
47—2 
