_Beta_ varies less than _Alpha_. Originally it consisted of a tall perpendicular line, on the right side of which four straight lines are so placed as to form two triangles, whereof the vertical line comprises the bases, while a small portion of that vertical line entirely separates the triangles ([Symbol]). This ungraceful figure was modified very early, even in inscriptions. On the Rosetta stone (No. 1) the triangles are rounded off into semicircles, and the lower end of the vertical curved. Yet the shape in manuscripts is not quite so elegant. The lower curve is usually the larger, and the curves rarely touch each other. Such are Codd. ANRZ, Rossanensis (sometimes), and the Cotton Genesis. In the Herculanean rolls the letter comes near the common cursive β; in some others (as Cod. Rossanensis at times) its shape is quite like the modern Β. When oblong letters became common, the top (e.g. in Cod. Bezae) and bottom extremities of the curve ran into straight lines, by way of return into the primitive shape (see No. 36, _dated_ A.D. 980). In the very early papyrus fragment of Hyperides it looks like the English R standing on a base (No. 9, l. 4). But this specimen rather belongs to the semi-cursive hand of common life, than to that of books.
_Gamma_ in its simplest form consists of two lines of equal thickness, the shorter so placed upon the longer, which is vertical, as to make one right angle with it on the right side. Thus we find it in the Rosetta stone, the papyrus of Hyperides, the Herculanean rolls, and very often in Cod. A. The next step was to make the horizontal line very thin, and to strengthen its extremity by a point, or knob, as in Codd. Ephraemi (No. 24), RZ: or the point was thus strengthened without thinning the line, e.g. Codd. Vatican., Rossanensis, N and most later copies, such as Harl. 5598 (No. 7) or its contemporary Parham 18 (No. 36). In Cod. Bezae (No. 42) _gamma_ much resembles the Latin r.
_Delta_ should be closely scrutinized. Its most ancient shape is an equilateral triangle, the sides being all of the same thickness ([symbol]). Cod. Claromontanus, though of the sixth century, is in this instance as simple as any: the Herculanean rolls, Codd. Vatican., Sinait., and the very old copy of the Pentateuch at Paris (Colbert) or “Cod. Sarravianus” and Leyden, much resemble it, only that sometimes the Herculanean sides are slightly curved, and the right descending stroke of Cod. Vatican, is thickened. In Cod. A begins a tendency to prolong the base on one or both sides, and to strengthen one or both ends by points. We see a little more of this in Cod. Rossanensis and in the palimpsest Homer of the fifth century, published by Cureton. The habit increases and gradually becomes confirmed in Codd. Ephraemi (No. 24), the Vatican Dio Cassius of the fifth or sixth century, in Cod. R, and particularly in N and E of the Acts (Nos. 4, 14, 25). In the oblong later uncials it becomes quite elaborate, e.g. Cod. B of the Apocalypse, or Nos. 7, 21, 36. On the Rosetta stone and in the Cod. Bezae the right side is produced beyond the triangle, and is produced and slightly curved in Hyperides, curved and strongly pointed in Cod. Z.