Self Bows and Composite BowsVegetius says that one quarter to one third of new infantry recruits, tirones, should be exercised as archers or sagittarii. He mentions wooden bows for training, suggesting that simple self-bows formed from one single piece of wood were used. It is not possible to be sure about the form of Roman self-bows. It is likely that Romans trained with long bows, similar to those found in Germanic lake deposits made from yew or fir. Different woods and materials are capable of storing varying amounts of energy, yew being the ideal wood for a bow. Some of these bows were self-nocked, with a second set of nocks cut into the stave. This has been interpreted as a way of shortening the string and increasing the power of the bow. But rather the outer set of knocks probably served for a stringer. Holding the bow horizontally the archer attaches a long length of string, the stringer, to the two outer nocks. By standing on the stringer and raising the bow to his chest a second man can slip the bow string over the inner nocks.
In combat the Roman military used the composite bow. This was based around a wooden core, upon which animal sinew and horn had been carefully glued producing a recurved shape. “Ears” were then attached to the ends of the bow, protected by bone or antler. These acted as levers for the flexible parts of the limbs. Laths on the grip stiffened the bow handle to prevent it flexing and bucking when the string returned to rest after release. Archaeologists have discovered these bone ear laths in several sites around Britain, Europe and Asia. To stop the glue breaking down in damp conditions, the limbs of the bow were waterproofed using bark and coated in lacquer. Recurved bows were produced for the Western Empire in fabricae near Pavia. In the East where such bows were more common the army could probably buy them on the open market. The recurved bows are shorter than the self-bows, allowing them to be used kneeling or on horseback. The materials used transfer more power from the bow to the arrow, making the bow more efficient.
The velocity of the arrow depends on the efficiency and power or draw weight of the bow, the draw distance from stave to ear normally being around 700mm, the weight of the arrow, and the drag of the arrow through the air. Penetration depends upon velocity, weight, the design of the arrowhead and composition of the target. Unlike medieval English archers taught from an early age, most Roman recruits would come to the discipline relatively late. While some medieval self-bows seem capable of achieving up to 180lb of pull, finds from votive lake deposits suggest 4th century bows would be less powerful. A pull of 80-100lb would seem reasonable for a Roman “D” section self-bow. After mastering the self-bow the archer could move on to a recurve bow of similar power, but greater efficiency.
The reconstruction of Roman bows cannot be an exact science. Little is known about the actual style of construction used for 4th century Roman recurved bows. The length of some of the bone and antler ear laths suggests the bows were not as heavily recurved as later Mongol examples, and asymetrical. A partially surviving Parthian weapon from Baghouz has been reconstructed. Known as the Yrzi bow after the area of the cemetery where it was found, the reconstruction had a draw weight of approximately 60-70lb. But the Middle-Eastern tradition of bow making would differ from that of the Huns of Central Asia, and possibly from whatever methods were used in the state fabricae. The Central Asian influence can be seen in the “Qum-Darya Bow”, found in a mass grave associated with a Chinese frontier post. Dated from the 1st century B.C. to the 3rd century A.D. similar ear and grip laths spread west with Hunnic influence. They are found in western Europe on such sites as Blucina and Wien-Simmering. Such bows used up to seven ear and grip laths, compared to four on the Yrzi bow.
Modern recurve bows cannot truly mirror those of the past. In the same way modern powerful yew self-bows do not exactly correlate to archaeological finds. My best 80lb bow is made from slow-grown North American yew, Taxus brevifolia. Grown in Oregon, it has a close grain without knots, and is an approximation of a Roman self-bow. European yew, Taxus baccata, would give a better cast and be more efficient but is rarely available today for making high poundage bows.
Two types of release can be used. The eastern or Mongolian release may have been used in the east from the 3rd century, with the arrow shot to the right of the bow and the string drawn back by the thumb. The eastern thumb draw uses a thumb ring, however a western or Mediterranean two or three-fingered release seems to have been the general method used in Roman Europe. Arrows were shot to the left of the bow, with the nock of the arrow held between the index and middle finger.
Arrows were either broad heads or trilobate designs to cut as much muscle and blood vessel as possible, or narrow bodkins designed to penetrate shields and armour. Arrowheads can be socketed or tanged. A shaft with a tanged head will be susceptible to breakage in some conditions. The weight of head will effect the distance the arrow can be shot. Light 40g flight arrows can travel over 200m from my 80lb bow. Indeed Vegetius states targets should be set up at 600 Roman feet (177m).
However it is probable that a selection of light, medium and heavy arrows would be carried to allow the archer to respond to a variety of threats. Heavy 70g arrows will have their greatest velocity during the first 50m of flight, and be used at relatively close ranges. Light arrows would allow a target to be peppered at 200m. The arrows from Dura and Micia in Roman Dacia were made of reed with wooden foreshafts to take the arrowhead. But arrows from bog burials give us versions of western-style arrow construction from solid pine or ash. Shafts had parallel sides, but were tapered slightly at the fletchings widening again at the knock to prevent breakage. This shape of shaft would also mean the arrows would leave the shaft cleanly. James suggested that the nature of the wooden foreshafts at Dura were not designed for holding a metal head, and the reed arrows were primarily for hunting or for target shooting. I have experimented with reed arrows and found them too weak to be shot from high poundage bows. They do seem better suited to weaker hunting bows.
Fletchings were not “corkscrewed” about the shaft to impart a spin. Such a system would only increase drag. The natural curvature and twist of the feather will impart a spin giving increased accuracy. Surviving fletchings from Dura show that western and eastern styles were used in the east. In the west a recognisable cock-feather lies at ninety degrees to the knock, in the east one of the three feathers is on the axis of the knock. Western style fletchings would also stop around 25mm from the knock to allow the fingers to grip the arrow. Eastern style fletchings can be taken to the end of the knock.