520 BOOMERANGS. 



of the bend, should follow the grain of the wood. When the projec- 

 tile falls hard upon one end the stress near the center is very severe, 

 and any point at which the direction of the grain meets the convex 

 edge obliquely is likely to develop a split and ultimately a breakage. 



It is better to cut the material to its final twisted shape rather than 

 to impart the twist Vjy another steaming and bending. Considerable 

 care is required in the process, for the removal of a layer of wood a 

 millimeter thick in such a way as to increase or diminish the twist will 

 cause a marked difference in the flight. It will be found to facilitate 

 throwing to cut that end of the boomerang which is held in the hand 

 to the somewhat square form shown at the right hand of figs. 1 and 3. 



There is some difficulty in avoiding warping, for boomerangs are 

 less likeh' to get broken if thrown when the ground is damp and soft, 

 and under these circumstances the moisture is likel}^ to be absorbed 

 by the wood. It is of great advantage, therefore, to make the surface 

 of the implements ver}- smooth with fine glass paper and to saturate 

 them with linseed oil. The additional densit}^ thereby produced is 

 also of service in that it diminishes the effect of the frictional resist- 

 ance of the air. 



I have used artificially bent oak as a material, but have not found it 

 as heavy or as strong as ash. Oak branches that are naturally bent 

 are not hard to procure, but boomerangs made from them are liable 

 to break at places where there are knots or irregularities in the grain 

 of the wood. 



EVOLUTION. 



Boomerangs of every variety of shape are still to be found in Aus- 

 tralia, and it appears impossible to get direct historical evidence as to 

 the nature of the successive stages of development. But if specula- 

 tion be allowed, the following series may be suggested: 



First, we should have a clumsy kind of wooden sword, curved, but 

 without rounding or twist, and with one end roughened to form a 

 handle; when the intended victim was out of reach it would be natural 

 to throw the weapon, and at short ranges it would be extremely 

 effective. Bad workmanship would involve the frequent production 

 of implements of which one side was more rounded than the other, 

 and it would soon be found that these missiles, when thrown with the 

 rounded side uppermost, traveled much farther and straighter than 

 the former. 



Boomerangs of this character vary in length from 50 to 110 cm., 

 and in weight from 200 to 1,250 grams. They are, for the most part, 

 twisted in a manner that seems quite fortuitous, and form the enor- 

 mous majorit}^ of the present native implements. Light specimens 

 with a slight left-handed twist may have a fairly straight trajec- 

 tory of 100 meters, and ma}^ return if aimed much uphill, especially 



