646 



SCIENCE PROGRESS 



slide for holding a light-filter. There is no optical system or heat-absorbing 

 cell interposed between the Pointolite lamp and the condenser, as these are 

 quite unnecessary with this source of light. No camera is used, as the 

 apparatus is always worked in a room illuminated with red light. The 

 plate is supported in a plate-holder rigidly fixed to the bench, usually about 

 one foot from the eyepiece of the microscope. • Focussing is done direct 

 onto a piece of white card already in the plate-holder. A simple, specially 

 designed vibrationless shutter is fixed to a support independent of the bench. 

 After focussing, this shutter is brought down just in front of the eyepiece 

 of the microscope. 



In practice a green light-filter is used with green sensitive plates. Green 

 light is used because the eye has the maximum sensitivity in this region, 

 and therefore focussing is facilitated. A process plate is used, because it 

 gives the best definition. The exposure required with a magnification of 

 2,000 diameters obtained by means of a 2 mm. oil-immersion objective 

 x 14 eyepiece is from 5 to 10 seconds. 



Octaval Notation for Inch Fractions 



A rule for engineers' precision measurements has just been issued which 

 embodies a proposal for a logical notation for the inch binary fractions. 

 Such notation is not new, but for the first time is put into a practical form 

 for the workshop. 



It is based on the observation that although (the tendency of compu- 

 tators being towards decimals) all micrometers and fine measuring instru- 

 ments have been made with decimal divisions of an inch, the instinct of the 

 craftsman or mechanic is towards a binary division into halves, quarters, 

 etc., down to 64ths. 



In practice the vast majority of English and American engineers stan- 

 dardise their machine parts, down to 64ths, in such binary fractions, 

 although for their measurements they have to use decimal fractions, no 

 other measuring notation having been provided. 



In place of the radix of 10 used for decimal notation, this rule (and pro- 

 posal notation for finer measurements) uses a radix of 8. 



Just as in decimals 0-234= f- J ^+-^ == 1 



J J h 10 io 2 io 3 10 100 



so in octavals, o 234 



1000 



?4_3__, £_. ? + l , JL 

 8 ^ 8 2 - 1 " 8 3 "' 8 T 64 T 512' 



The rule here illustrated divides the inch into 8 equal parts numbered 

 1 to 8, and each of these parts is sub-divided (by stepped divisions for the 

 even values) into 8 parts again, making the finest divisions 64ths. The 

 " staircase " divisions make it easy to know the value of each sub-division 

 by its length. 



The " pip " which takes the place of the decimal " point," is the printer's 

 degree mark upside down, and obviates confusion with decimal fractions. 



With this notation all binary fractions down to 64ths are expressed in 

 two figures accurately. A table of these octaval equivalents, and some 



