LXVIII,. T. H. HOUGHTON. 
The method of construction adopted by Mr. Airey is to attach 
a screw of one, two, or three threads formed on a core, to a cylin- 
drical casing, and revolve the whole, in the older types the screw 
simply revolved in a stone or timber trough, rather higher than 
the centre line of the worm, a considerable leakage takes place 
between the screw and the trough, and although cheaper to con- 
struct, the efficiency is less. Figure 4, shews one similar to the 
screws used at the Antwerp Water Works for raising water into 
the filter beds. They are not suitable for a variable height in the 
delivery channel, as the outlet of the pump must be above the 
highest level, and any fall in the level of the water in it, reduces 
the efficiency of the machine, through the water being raised to a 
greater height than is required by the difference between the top 
and bottom water levels. 
Chain pumps are largely used for raising small quantities of 
water, and when made with buckets that are a good fit in the 
barrel give fairly good results, for large quantities of water, say 
1000 gallons per minute, the diameter of the rising barrel would 
be great, as the speed at which the buckets can be raised is not 
more than two feet a second, the efficiency of this class of pump, 
if well constructed is about 55°/, but its use is limited, both as to 
quantity and height to which the water has to be raised. With 
a varying height in the delivery channel the efficiency becomes 
less as the level falls. 
Rotary pumps have been used for pumping small quantities of 
water for some years, but the author is not aware of any at work 
on a large scale, some large ones were erected about ten years ago, 
at one of the London Water Works for pumping water on to the 
filter beds, they proved a failure, and were removed by the con- 
tractors, three throw pumps being erected in their place. One of 
the many makes of this class of pump is shewn in Figure 5, @ 
great difficulty is experienced in keeping the pistons tight against 
the sides of the casing, and as a consequence there is either undue 
friction, or leakage. Mr. Isherwood U.S. Navy records in the 
Journal of the Franklin Institute for 1889, some tests made on 
