Feb. 29, 1 8 72 J 



NATURE 



345 



the principal element in Pere Secchi's computations of 

 solar temperature. It will be seen on referring to the 

 second and third columns of the table that, while the 

 upper thermometer indicates a mean temperature of S6'9,° 

 the lower one shows only 79'5°, difference = 7"4°. This 

 great discrepancy of temperature at different points of the 

 upper portions of the annular space at which, owing to 

 the inclined position of the concentric tubes, something 

 like uniformity ought to exist, suggests a still greater dis- 

 crepancy of temperature at the underside towards the 

 lower termination of the tubes. In addition therefore to 

 the observed irregularity of temperature at the upper part, 

 shown by the table, no indication whatever is furnished of 

 the temperature of the fluid in the annular space below the 

 central lube, nor towards the termination at either side. 

 Obvious!)', then, no accurate computation can be made of 

 the degree of refrigeration to which the central thermo- 

 meter is exposed by the radiation from the cold blackened 

 surface of the internal tube, every part of which, as we 

 have seen, possesses a different temperature compared 

 with the rest, consecjuently transmitting radiant energy of 

 different intensity. It will be found practically im- 

 possible, therefore, to determine the true differential 

 temperatuie of the contents of the bulb exposed to 

 the sun's rays and the fluid contained in the annular 

 space. Hence, the differential temperature entered in 

 the table, the result of comparing the indications of the 

 thermometers, is manifestly incorrect. It will be found 

 also by reference to the table that while the mean tem- 

 perature imparted to the central thermometer by the sun's 

 rays is 93'i°, the mean temperature of the fluid in the 

 annular space is S3'3°. Consequently, the intensity of 

 solar radiation established by the instrument is only 

 93'i° — 83 3° = 979° Fah. Now, the sun during the 

 recorded experiment of September 2 was exceptionally 

 clear, the mean indication of the actinometer while the 

 experiment lasted being 6o'05°, thus showing that the 



energy developed was only 9-Z2- = o'i6 of the true radiant 

 6005 



intensity. The mean zenith distance, it may be men- 

 tioned, was only 33' 24' during the experiment. Agree- 

 able to the table of temperatures previously published, the 

 maximum solar intensity for the stated zenith distance is 

 63'35' ; thus we find that the sun, as stated, was excep- 

 tionally clear while the trial took place, which resulted in 

 developing the trifling intensity of g'yg" Fah. The result 

 of the experiments conducted September 6th, recorded in 

 the table, it will be seen was nearly the same as that just 

 related, the mean temperature indicated by the thermo- 

 meter exposed to the sun being 9S'2^, while the mean of 

 the two thermometers immersed in the fluid was 8y8°, 

 hence the differential temperature 98'2'' — 87'8°=lo'4°. 

 The mean temperature of solar radiation during the ex- 

 periment, ascertained by the actinometer, was 59■75^ the 

 zenith distance being 35° 33'. Consequently, the intensity 



indicated September 6th was only — 4^ = o'i7 of the true 



5975 

 energy of the sun's radiant heat, against o'i6 during the 

 previous experiment. It will be observed that the fluctua- 

 tion of the differential temperature was much greater 

 September 2nd than during the succeeding experiment, 

 owing, no doubt, to the influence of currents of air pro- 

 duced by a strong breeze on the first occasion, the re- 

 volving observatory being partially open on the side pre- 

 sented to the sun during observations. 



With reference to the small differential temperature 

 indicated by the Secchi instrument manufactured by 

 Casella, it may be urged that it is not intended to show 

 the true intensity of solar radiation on the earth's surface, 

 but simply a means of determining solar temperature. 

 Granted that such is the object, yet the extreme irregu- 

 larity of the temperature of the fluid within the annular 

 space shgws that the instrument is unreliable, a fact 



established beyond contradiction by an experiment in- 

 stituted September 27, 1871. On this occasion water of 

 a uniform temperature was circulated through the annular 

 space. This was effected by gradually charging this space 

 from the top, and carrying off the waste at the bottom, 

 holes having been drilled in the external casing for thatpur- 

 pose. The result of this conclusive experiment is recorded 

 at the foot of Table A. It will be found on reference to the 

 figures, that the mean difference of the two thermometers 

 immersed in the fluid was only 64 '9°- 64^4 = o'5^, while 

 the mean differential temperature was augmented to 

 79'i° - 64-45 = I4'65° against 979° on the 2nd of Septem- 

 ber, although the zenith distance was greater, and the 

 solar intensity less ; circumstances which ought to have 

 diminislied the indicated intensity. It is needless to enter 

 into any further discussion of the demerits of the instru- 

 ment represented in Fig. 2. We may now return to the 

 consideration of the device delineated in Fig. i, copied 

 from " Le Soleil." It will be seen that the material 

 difference of construction is that of applying only one 

 thermometer for ascertaining the temperature of the fluid 

 in the annular space. Possibly this single thermometer 

 may indicate approximately the mean temperature of the 

 upper and lower portions of the fluid above the central 

 tube ; but it furnishes no indication of the temperature 

 below, nor at either extremity of the annular space. The 

 inadequacy of the means adopted for ascertaining the 

 temperature of the internal surface which radiates towards 

 the bulb of the central thermometer having thus been 

 pointed out, it will be well to consider whether the ex- 

 pedient of passing a stream of water of nearly uniform 

 temperature through the annular space, will insure trust- 

 worthy indication. In order to determine this question, 

 I have constructed two instruments, in strict accordance 

 with the delineation in Fig. 1, excepting that in one of 

 these the concentric cylinders are considerably enlarged, 

 the annular space, however, remaining unchanged. Ex- 

 periments with the two instruments prove that the enlarge- 

 ment does not materially influence the indications, 

 provided water of a uniform temperature be circulated 

 through the annular space. But these experiments have 

 demonstrated that the size of the bulb of the thermometer 

 exposed to the sun cannot be changed without influencing 

 the differential temperature most materially. This will 

 be seen by reference to Table B, which records the result 

 of experiments with different thermometers, and tubes of 

 different diameter, conducted October 17, 1871. As on 

 previous occasions, the instruments, in order to insure 

 accurate position, were attached to the dechnation table 

 arranged within the revolving observatory. The bulbs of 

 the thermometers employed were very nearly spherical, 

 their diameters being respectively 030 and o'58 ins. 

 The upper division of Table B which records the experi- 

 ment with the small bulb exposed to the sun, establishes, 

 it will be seen, a differential temperature of I4'4^ for the 

 instrument having the i}-in. central tube, and 16° for the 

 one having the 3-in. central tube. Referring to the 

 lower division of the same table, it will be seen that 

 when the thermometer with the lat-gc bulb is exposed to 

 the sun, the differential temperature reaches 22'5° in the 

 instrument containing the ijin. central tube, and 2ri"' in 

 the one having the 3in. tube. We thus find that, by 

 doubUng the diameter of the bulb of the thermometer 

 exposed to the sun, all other things remaining unchanged, 

 an augmentation of the differential temperature amounting 

 to nearly one-third takes place. This fact proves the 

 existence of inherent defects fatal to the device delineated 

 in Fig. I, rendering the same wholly unreliable. 



Agreeably to the doctrine of exchanges, the diameter of 

 the bulb is an element of no moment, since the internal 

 radiation towards the same — provided its teinpcratiire be 

 uin/or/)i^depends solely on the temperature and angular 

 distances of the radiating points of the enclosure. In- 

 fallibility of the " solar intensity apparatus " has evidently 



