1907-8.] On the Theory of the Leaking Microbarograph. 453 
In making time-marks on the microbarogram, it is desirable to avoid 
handling the instrument. All that is necessary is to shut the door of the 
room containing it rather briskly at a known time. This will cause the pen 
to draw a short, sharp line across the trace, and the time can be written 
opposite these marks when the chart is removed from the drum. 
There are two other respects in which the construction of the instrument 
should be improved. A more refined arrangement should be introduced 
to secure that the point at which the hook connecting the stem of the 
floating plunger to the axis of the recording lever touches this axis 
does not shift arbitrarily ; so that the pen shall return exactly to the same 
place after the instrument has received a jar or the air-pressure has 
been suddenly altered by slamming the door of the observing room. 
Also, the steadying spring and all other sources of friction should be 
removed from the stem of the plunger, or at least be removable, when it 
is at work. 
During the observations the leakage of the Lochearnhead instrument 
increased, so that it came to register only the more transient or more 
violent disturbances. Thus on many interesting occasions its readings 
were unfortunately of no use. For good work it is essential that the 
microbarograph should be so constructed that it can be rapidly taken to 
pieces, tested, adjusted, and fitted up again. Considering the great 
simplicity of its general design, this should not be difficult. 
As regards the nature of the phenomena under observation, the follow- 
ing points are to be noted : — 
A set of observations from three stations only affords no means of 
testing the admissibilit}^ of our assumptions regarding the rectilinearity of 
wave-front and path of propagation. The observations themselves seem to 
indicate that the breadth (in the direction of propagation) of the dis- 
turbances in question is not very great, say 10 to 40 miles. Also there 
were cases in which a disturbance was not observed at all the three 
stations. It is therefore unlikely that the assumption of a rectilinear wave- 
front can be more than a very rough approximation. 
Again, making all allowance for want of absolute similarity in the 
instruments, there seems to be no doubt whatever that the distribution of 
pressure in the disturbance varies as it progresses. This introduces un- 
certainty in identifying the points on the microbarogram which correspond 
to the same phase of the disturbance. We are, in fact, face to face with a 
wave which changes form as it goes along, and therefore has no definite 
velocity of propagation in the ordinary sense. 
All these qualifications must be borne in mind in accepting the follow- 
