MKASfRING OF A IK. 433 



Measurements of temperature were in all cases made with 

 ordinary mercury in glass thermometers and these have not been 

 standardised — no exceptionally high temperatures occur in this 

 work — but there can be no gross errors as most tem^K-ratures 

 were taken in duplicate. 



The amount of air passing through the compressor was 

 measured on the intake as well as on the discharge side. As the 

 compressor has two low pressure cylinders, working in parallel, 

 two oritices of the same design had to be used for the intake 

 side. These orifices are very similar to the upstream portion of 

 a Venturi tube. By comparison with the standard capacity 

 meter at Ferreira Mines the coefficient of contraction for \ enturi 

 tubes was found to be 0.985 and, therefore, this coeiiicient has 

 been taken as the most probable value for this type of orifices. 



The value of A for the intake orifices has been taken as 

 infinity though the floor level was only about 2' 6" below the 

 mouth of the orifice and there was a wall about 15" from the 

 hne joining the centres of the two orifices. 



It may be considered correct to treat the orifices as though 

 they were situated in free space. 



Besides having a proper approach for the air to these orifices, 

 it is also necessary to have an unintierrupted discharge on the 

 downstream side of the orifice. 



The arrangement of the butterfly valve in the downstream 

 side of these orifices is not ideal, and involves a certain inter- 

 ference in the stream lines which showed in the readings of the 

 water columns which were used for measuring the drop of 

 pressure across the intake orifices. 



On account of the comparatively large area of the intake 

 pipes, four water manometers were used for each orifice. 



Although tihese four manometers were arranged symmetric- 

 ally with regard to the orifice they could not be symmetrical with 

 regard to the butterfly valve 'and this want of symmetry was 

 reflected in the readings of the water columns. 



The four water columns at each orifice were of slightly 

 difl'erent height, so that it would have been necessary to use four 

 observers per orifice in order to obtain a correct observation. 

 It was, however, not feasible to put four observers to each orifice 

 and the dit^culty was overcome in the following way. 



The four pipes were fixed on a board close together and the 

 observer found the a\erage height of the four water columns by 

 means of a glass plate on which a dark line was cut, by following 

 closely the movement of the water columns. 



