^{page 155 note 1} However there are earlier examples of transition style. In the library of Etchmiadzin No. 102/2679 is the earliest Armenian paper MS. known, dated 971. A vellum Armenian MS. fragment is dated 1040. The MS. No. 1687/1890 is dated 1174, MS. No. 921/948 dated 1196, MS. No. 163/2606 dated 1198 are all on paper, written by different scribes, and have great similarity of writing with my MS. There are some others in the transition style in the same library, No. 1561/1568 on vellum dated 1173, No. 1239/1214 on paper dated 1118, No. 2093/2101 on paper dated 1223, No. 482/3795 on paper dated 1190.

^{page 156 note 1} The only other Kozma known to me is an Armenian MS. illuminator active 1662–1695 in Eastern Armenia. Kozma's mother's name Maria is definitely un-Armenian, and in Greek form.

^{page 156 note 2} For a brief description of my MS. see Professor Allen Wikgren's article, “Armenian Gospel MSS. in the Kurdian Collection,” MS. No. 2, in the Journal of Biblical Literature, lv, part ii, 1936, pp. 155–8Google Scholar.

^{page 157 note 1} At the present No. 11 of Four Gospels of my collection of MSS.

^{page 159 note 1} I am deeply grateful to my great friend Dr. Harold R. Wfflough by of the University of Chicago for communicating this information to me. He has done much excellent work in Greek Iconography.

^{page 159 note 2} See my article, “Miniaturist Iknadios,” with twenty-six illustrations of his miniatures in the Armenian periodical Anahid, No. 3, Paris, France, 1939Google Scholar.

^{page 161 note 1} As basis for our assumption 216 Armenian Four Gospel MSS. dated from the twelfth to the end of the fifteenth century were taken into consideration. These are catalogued in The Catalogue of Armenian MSS. of Vassbouragwn, by Lalaian, Yervant, vol. i, published in Tiflis, 1915Google Scholar.

Some of these have been restored a second time, i.e. Nos. 47 in 1559, 51 in 1741, 167 in 1668, 238 in 1792, 142 in 1810. Beside these No. 63 written in 1294 restored in 1386 and again in 1580. No. 68 written in the thirteenth century restored in 1503 and again in 1592.

In this group of forty-eight MSS. only four have been restored in periods of less than one hundred years. Thus less than 8 per cent of the MSS. are restored or need restoration during the first hundred years of their existence. No. 77 written in 1304 was restored in 1374; however, this is a rare example.

Also in this group we have seven MSS. that have been restored a second time. No. 167, written in 1444, was restored in 1612 and again fifty-six years after in 1668. The rest were restored a second time not less than 90 years after their first restoration, and one 275 years after.

Thus our assumption based on these facts has justification when we accept one hundred years as the time that elapsed between restorations.