148 BULLETIN OF THE ESSEX INSTITUTE. 



inclination and the uniform length of the object (the average 

 deviation from the mean being very slight) , the evidence 

 that the object has a front and back, and an above and be- 

 low, the spines springing at right angles to the plane of 

 the rings. With these constants, so to speak, might be 

 added the importance of those surfaces showing signs of 

 wear, as well as the bending of the spines and manner of 

 breakage. 



The bow-pullers are found associated with Etruscan, 

 Early Roman and Early Greek remains ; they belong to 

 pre-classic and early classic times. They have been found 

 in the tombs of warriors. In two examples figured by 

 Strobel, chains with large 8-shaped links are drawn through 

 them (Fig. 10). It was the association of the chain in 

 this way that led Strobel to conceive the object to have 

 been designed for some form of snaffle or curb for horses. 

 Reference to Strobel's memoir will be made further on. 



While in nearly every instance this object is labelled in 

 museums bow-stretcher or bow-puller, authorities have 

 not fully accepted this Interpretation without question. 

 Gozzadini in his memoir on the ancient Etruscan Cemetery 

 of Marzabotto near Bologna says : " Archaeologists have 

 agreed, but I do not know on what foundation, in suppos- 

 ing that certain double rings provided with three points 

 were used by archers by inserting the middle and fore fin- 

 ger in order to stretch the cord. Now this attribution is 

 strengthened by Tommsen, Director of the Museum at 

 Copenhagen, who told Cavedoni that some ' of these im- 

 plements were found placed together with bronze bows in 

 caves in the northern countries.' They are found in all 

 Egyptian, Etruscan and Roman Museums and they are 

 taken out of Lacustrian stations, and Torbiere (Pit 

 graves ?) and they find them again in opening the Necrop- 

 olis of Marzabotto without, however, that phallic sign 



