At the outset I met with a very curious and unaccountable discrepancy in the form of release employed, and that was when the archer was represented with his right side, or shaft hand, toward the observer, the hand was with few exceptions in the attitude of the primary or secondary release; whereas if the archer was represented with his left side, or bow hand, toward the observer, the release with few exceptions represented the Mediterranean release. Or,, in other words, as one faces the sculptured slab the archers, who are represented as shooting towards the right, show with few exceptions either the primary or secondary release, while those shooting towards the left are with few exceptions practicing the Mediterranean release!

If in every case the Assyrians were represented on the left, as one faces the tablet, fighting the enemy on the right, then one might assume that the enemy was practicing a different release. In an Egyptian fresco, for example, where Rameses II. is depicted in his chariot fighting the Arabs, the enemy is represented as practicing a different release. While in many cases the Assyrians are on the left of the picture, in other cases they are on the right, and shooting towards the left. It is therefore difficult to decide which release was practiced by them ; and all the more so, since, with very few exceptions, the releases are perfect representations of forms practiced today, which have already been described. I have suspected that in one or two cases the Mongolian release might have been intended, though in no case is the thumb-ring represented, though other details of arm-guards, bracelets, etc., are shown with great minuteness.

Taking the releases as they are represented in the sculptures without regard to the discrepancies above noted, it is an extremely interesting fact that all the earlier Assyrian archers, that is, of the time of Assurnazirpal, or 884 B.C., the release represented is the primary one, as shown in Fig. 21 ; while in the archers of the reign of Assurbar-nipal, or 650 B. C, the secondary release is shown, or a variety of it, in which the tips of all three fingers are on the string, as shown in Fig. 22. Between these two epochs the sculptures ranging from 745-705 B. C, notably a slab representing the campaign of Sennacherib showing assault on the Kouyunjik Palace, both the primary and secondary releases are represented. If any reliance can be placed on the accuracy of these figures, an interesting relation is shown in the development of the secondary from the primary release, as urged in the, first part of this paper. Possibly a proof that the primary release is intended is shown in the fact that the arrows are represented with the nock end bulbous.

On tablets in the British Museum of this intermediate age, or during the reign of Tiglath Pileser, is the first representation of an archer with the right side towards the observer practicing the Mediterranean release : and on slabs of the date of 650 B. C, one showing Assurbarni-pal's second war against Elam, and another one representing the siege of the city of Al-ammu, a number of archers with their right towards the observer are practicing the Mediterranean release (Fig. 23). In the Mediterranean release, which, as I have before remarked, is represented, with few exceptions, by all the archers having the bow-hand towards the observer, there are two varieties shown; one in which three fingers are on the string, and another with only two fingers drawing the bow, as shown in the accompanying figures (Figs. 24, 25). The Mediterranean release occurs in Assyrian sculpture as early as 884 B. C, as shown on a marble slab in the British Museum representing the siege of a city by Assurnazirpal (Fig. 26). A curious form is shown in Fig. 27, showing Assur barnipal in a chariot, shooting lions. The string below is concealed by the archer's arm. The secondary release is probably intended.