A HISTORY OF HEREFORDSHIRE used for other purposes ; while on the other hand, where they were deficient or absent, the lord must have entrusted his plough-oxen to the hands of some other class."' There is some evidence that in the Conquest period the serfs were diminishing in number,"" while in Essex, as the present writer has shown, there was a marked increase in the bordars, the class intermediate between the villein and the serf. It is only where we are given the figures TR.E. and T.R.W. that we can make the comparison, and in Herefordshire we have these figures for the notable manor of Leominster. The Domesday figures are these : — Villani Bordarii Servi et Ancillae T.R.E 238 75 82 T.R.W 224 81 2S At the same time the ploughs in demesne had been reduced from 30 to 29, and those of the peasants from 230 to 201. There is nothing here to account for the great decrease of serfs, and as there were no bovarii, the serfs remain- ing in 1086 were far too few in number for the lord's demesne ploughs. To whom, then, were his oxen entrusted ? The same question is suggested at Hayles in Gloucestershire to the south. There were twelve serfs on that manor TR.E., but none in 1086, because the owner had freed them. And yet there were three ploughs on his demesne. By whom were they worked ? Professor Tait guardedly suggests that perhaps bovarii took their place."^ But there were no bovarii at Hayles and none at Leominster. At the latter place we have only a small increase in bordarii, the same class who are found, we saw, to have increased in Essex. The proportion of serfs to the recorded population has been worked out by Seebohm for each county, but only from EUis's figures ; and ElHs's figures are here peculiarly in need of emendation."* But even as they are they show clearly that the serfs, here as elsewhere, were much in excess of the ancillae. But in Shropshire the excess was far greater, and the dispro- portion in Gloucestershire was relatively enormous. The low percentage of the serfs was accountable for this rather than a high one of ancillae. On the lands of the church of Hereford this local feature was accentuated. Indeed we have actually eight entries, in each of which there is one ancilla, and one in which there are two, but no servi at all. In another there is one of each class ; in three there is one servus to two ancillae ; and in one there are four to six. The total, it is true, of serfs for the fief exceeds that of the ancillae, but the above facts are very noteworthy, as is the large number of estates on the fief on which there were no serfs. We are much in the dark about the ancillae, though they are recognized as a servile class ; one cannot assume that, when entered alone, the ancilla was in fact the dairymaid, for that '" Prof. Maitland guardedly suggested that the servi may have been entered on different principles in different counties and have existed in some where none are mentioned, but this would hardly affect the above argument. '" Dom. Bk. and Beyond, 35 ; F.C.H. Essex, , 359-63. '" ' The statistics given in the Essex Domesday show that in that county there was a large decrease in the number of the servi within twenty years of the Conquest. The " bovarii " of the west may conceivably be the result of a similar movement there' (F.C.H. Shrops. i, 303). '" Ellis reckoned 69 1 servi and 99 ancillae, but we have first to add all the entries of servi and ancillae entered jointly, as at Leominster, where there were 25, and Cobwell where there were 9, then to deduct, for manors not in the county (fol. liob), 12 servi, 5 ancillae, and 28 servi and ancillae jointly, and finally to'add the figures for the manors now in Herefordshire, which were then entered under Shropshire, as Leintwardine (where there were 5 servi). 290
