The determinations of relative hardness were made by comparing each of the woods 

 with those forming the following scale : 



1. Erythrina caffra (Kafirboom), 



2. Pinus picea (Pinewood of N.G.R. sleepers). 



3. Odina Caffra (Kafir Plum}. 



4. Royena lucida (Zwartbasf). 



5. Brachyama discolor (um-pahla). 



6. Elaeodendron velutinum (um-Nqai). 



7. Olea foveolata (Bastard Ironwood). 



8. Pteroxylon utile (Sneezewood). 



9. Milletia Caffra (um-Simbiti). 

 10. Red Ivory. 



The specimens forming this scale were so selected as to make the corresponding 

 numbers represent, approximately, the relative resistance to small indentations of uniform 

 size. 



TRANSVERSE TESTS. 



The transverse tests were made with a simple form of lever-testing machine con- 

 structed for the purpose at the Railway workshops. The specimens used were planed 

 exactly 4 centimeters square, and cut of a sufficient length to give the necessary bearing 

 on the supports ; these were shod with iron plates slightly rounded on the edge and set 1 

 meter apart. 



The loads were applied carefully and gradually by means of a screw, and transmitted 

 by an interposed lever to a rod, 12 millimeters in radius, guided horizontally by a slide ex- 

 actly midway between the supports. The pressure at any moment was registered by spring 

 balances, and the deflections by a scale on the slide. The deflections were read for 

 successive loads of 50 kilograms. When the load reached 200 kilograms, it was removed 

 and a reading of the set taken. Pressure was again applied and the deflections, for loads 

 of 250 kilograms and more, read, until the specimen failed. When the limit of elasticity 

 was approached or exceeded, the deflections were not read before they ceased to increase 

 sensibly with time. 



W I 3 

 The formula used in calculating the coefficient of elasticity was E=- , I, b, 



h, being the length, breadth, and depth of the bar in millimeters, W, the load in 

 kilograms. The deflection A selected, was that for a load of 100 kilograms, except in 

 cases when the limit of elasticity was reached with a smaller load. 



The formula used in calculating the modulus of rupture was R = , /2 , I, b, h, 



being taken in centimeters, and W, in kilograms. 



The diagrams in Plate V. show the behaviour of some of the principal Natal woods 

 under transverse pressure. The loads applied are indicated by the ordinates from each 

 small circle, and the deflections by the corresponding abcissae, a different origin being 

 used for each curve. When the limit of elasticity is exceeded, the deflections cease to be 

 proportional to the loads, and the curve departs from an approximately straight line. The 

 last test within the limit of elasticity is shown by a larger circle. These diagrams show 

 that the elasticity of the different specimens varies considerably less than their resistance 

 to rupture. 



CRUSHING TESTS, 



The crushing tests were made by means of a 20-ton Tangye's hydraulic testing 

 maehine, on specimens 4 centimetres square and 32 centimeters long (8 diameters). The 

 test pieces were placed horizontally between the platforms of the machine, and pressure 

 was applied until they failed. The pressure was registered by a manometer. The 

 results are given in Table III. 



GENERAL. 



The properties of each kind of wood vary considerably with the density and length 

 of seasoning : 



a. The hardness increases with the density and with the length of seasoning. 



6. The strength, both crushing and transverse, increases nearly in proportion to tha 

 density and by thorough seasoning, it may be increased by one-fourth. 



c. The coefficient of elasticity increases with the density, but only slightly with the 

 length of seasoning. 



d. For the same sectional dimensions, the greater the length, the smaller the crushing 

 stress and the modulus of rupture. 



