460 Proceedings of Royal Society of Edinburgh. [sess. 
same transparency and refrangibility as tbe liquid itself. What- 
ever may be the motions of these bodies within the liquid, they 
can have no disturbing effect on the course of the rays passing 
through the medium, which will behave as an homogeneous 
substance. 
This reasoning leads us to expect the most perfect telescopic 
images whenever the atmosphere traversed by the light of the 
star is in the state of adiabatic equilibrium. Now, it is a well- 
known fact that this state is reached, or at least approached, when 
air is agitated by convection. It is for this reason that Lord 
Kelvin long ago proposed to call this equilibrium ‘convective,’ 
instead of ‘adiabatic’ or ‘indifferent.’ Hence we conclude that 
seeing should be most favourable when the air has been previously 
stirred by convection-currents. With regard to the general 
atmosphere, we reach therefore the same conclusion at which 
Langley has arrived by his experiments where he considered the 
comparatively small mass of air in the immediate vicinity of the 
instrument. 
Several facts may be mentioned which seem to corroborate this 
explanation, and in some measure to bear out its validity. For 
instance, we know that on clear summer days, especially at 
continental stations, convection between the upper and lower 
layers of the atmosphere takes place during the daytime, being 
most energetic in the afternoon. Hence we infer that convective 
equilibrium is most nearly attained in the early evening, and 
consequently that the definition of stellar images should be best 
during the first hours of the night. In the later hours the seeing 
must become worse, because, in consequence of nocturnal radiation, 
the vertical distribution of temperature changes gradually so as 
to become incompatible with the conditions of adiabatic equilibrium. 
Towards the morning hours conditions become, therefore, more 
and more prevalent under which spontaneous displacements of 
masses of air must be accompanied by fitful changes of its re- 
frangibility. My experience as an observer at Strasbourg is in 
perfect accordance with these conclusions. As a rule, the seeing 
in the early summer evenings at the time of sunset was excellent, 
while after two o’clock in the morning the images had usually 
become so bad that the observations had to be discontinued. 
