512 DEDUCTIONS FROM SOME FIELD ARTILLERY PRACTICE. 
That is to say we ought to have some such pattern as this with our 
percussion Shrapnel: — 
2000 2025 2050 2075 2100 
+ 
+ 
+ 
where there are 3 — ’s within 51 yards of the elevation 2050, and also 
3 +’s within the same distance of the same elevation. 
As special cases, V for the 200 yards bracket may be considered 
equal V for the first 50 yards bracket as J l / 2 . 
Thus supposing we had a 200 yards bracket at about 1000 yards 
range, we should have 
(for time effect) Sh — ^ * _I — 1 5 
(for effect per shell) Sh — — - — ^ 
For Time EU'eet. 
V 
xooo yards. 
2000 yards 
3000 yards. 
4000 yards. 
£ 
15 Rounds. 
xo Rounds. 
9 Rounds. 
7 Rounds. 
1 
30 „ 
21 ,, 
*7 „ 
IS 
i£ 
45 „ 
32 „ 
26 „ 
» 
2 
Co „ 
42 „ 
35 ,, 
30 
3 
90 
64 „ 
53 „ 
45 
4 
120 ,, 
85 
69 
Co „ 
5 
150 
106 ,, 
86 
75 
The above table allows for ordinary misobservation. But when mis- 
observation is 50% and there is no range-finder, the only course, if or¬ 
dered to fire, is to do so in the general direction of the target without 
any ranging or verification at all. Similarly the analyses appear to 
show that, if the range is measured once and if the misobservation is 
likely to be 30%, it is best to trust entirely to the range-taking. The 
average range-taking error is about 3%. Should it be much less or 
much more, the above remarks must be modified. Every time the 
range can be measured to different parts of the target by different 
observers, is a$s distinct a gain as is an extra 200 yards bracket, or 
somewhere thereabouts at medium ranges, provided always that cor¬ 
rections for weather, powder and displacement are not entirely lost 
sight of. If these points are really carefully attended to, the instru¬ 
ment has a higher value. 
The table given above for verification shows that, when the allow¬ 
ance of time or of ammunition is very limited, one round may change 
the value of V by 100%. To keep exactly to what the table prescribes 
is therefore clearly impossible at practice. That, however, is a matter 
of little moment, for, first the table is very rough, secondly there are 
almost always two values of V that give equal results and thirdly there 
