PROFESSOR TAIT ON MIRAGE. | 567 
19. We have another and independent mode of testing whether this value of 
é accords with observation. For Scoressy tells us that, only on rare occasions 
and then only slightly, were objects at four miles’ distance affected. The usual 
distance was 10 to 15miles. Now, by the table in § 15 we see that the nearest 
object of which an image can be formed is distant 
ae T-Gfeck: 
uO 
or, with the above value of e, about 12 miles. 
There is thus a fair agreement, so far at least as these tests can tell us, 
between the results of our hypothesis and observation. 
The table in § 15 shows that, with the same value of e, and the same thick- 
ness of the lower stratum, as before, but with the assumption of a transition 
stratum of a thickness of five feet only; the distance of the nearest object of 
which an image could be formed would be about six miles only. A still farther 
reduction of the thickness of the transition stratum reduces this least distance 
still farther; but it is clear from the table that there is a limit somewhere 
about five miles. This would be still farther reduced if we supposed the lower 
uniform stratum to have a depth of less than 50 feet. On the other hand, we 
see that an increase of thickness of the transition stratum introduces distances 
greater than are consistent with observation; unless indeed, the thickness of 
the lower stratum be at the same time reduced. In the table # and @ depend 
upon the ratio of 6 to ¢; € is proportional to c. 
20. The columns headed & in the table of § 15 give, as shown in § 12, the 
magnitudes of the images relative to that of the object seen directly. They 
show that the inverted image is always taller than the object. This is consistent 
with Scoressy’s observations. When the object is not near the critical 
distance, however, this magnification is not considerable :—even if we assume 
a 50-foot transition stratum. On the other hand, the erect image, except when 
the object is not far beyond the critical distance, is much smaller than the 
object. Moreover, as is obvious from §§ 12, 15, this image is seen by converging 
rays. No doubt they are so nearly parallel as to be capable of producing 
distinct vision in a normal eye; but the remark is necessary as showing how 
different, in some respects, is the phenomenon from one of WOLLASTON’Ss imita- 
tions of it. Both images become infinite :—7.e., there is simply ‘“ looming ”:— 
when the object is situated at the critical distance. And, as the tables show 
from the result of § 13, the ratio of the distance between the images to the 
apparent size of the object seen directly, increases as the object recedes 
beyond the critical distance. All this seems to accord completely with V1ncE’s 
and ScoREsBY’s observations. The only additional remark I need make is that 
VOL. XXX. PART II. 4T 
