June^, 1878] 



NATURE 



153 



Similar results are obtained for the other quarter luna- 

 tions. In all cases that oscillation is the greatest of the 

 two for which the sun is above the horizon, whether the 

 moon be above it or not. 



There are still some remarkable facts connected with 

 this variation at the magnetic equator. Limiting our 

 examination of them always to December and January, 

 we find, if we determine the oscillations due to the moon 

 for the day when she is in conjunction and for each of the 

 six following days, that in the first three days of the seven 

 the oscillation is west-east-west during the day, that is, 

 from sunrise to sunset ; and in the last three days it is 

 east-west-east. In the middle day of the seven the lunar 

 action is almost null ; the oscillation of the needle is very 

 small, as we might expect, since on that day the change 

 at sunrise from a west-east to an east-west motion takes 

 place. The lunar hours of the maximum and minimum 

 extremes thus oscillate about two hours on each side of 

 the mean, depending on the position of the moon at 

 sunrise. 



The action of the moon, then, is dependent on the 

 sun's position relatively to the equator (or the earth's 

 position in its orbit), and on the position of the moon 

 relatively to sunrise and sunset. But there is no rela- 

 tion between the laws and amplitudes of the solar and 

 lunar diurnal oscillations. In the months from which I 

 have taken my illustrations, the solar diurnal variation 

 is a single oscillation : that for the moon, however taken, 

 for single days, for quarter or for whole lunations, is 

 always double. Through the combination of all the 

 varying modes in which this oscillation is produced from 

 day to day, the mean for a lunation is a regular double 

 oscillation. The amplitude of this mean oscillation is 

 three times as great in January as in June or July ; 

 whereas the amplitude of the mean solar diurnal varia- 

 tion is a half greater in June or July than in January. 



I shall add another fact, one of the greatest im- 

 portance in connection with this subject. We have seen 

 that the lunar diurnal variation changes in the relative 

 amplitudes of the two oscillations from day to day ; the 

 consequence of this is that when the means for a whole 

 lunation, or even a quarter lunation, are taken, the mean 

 amplitude is much less than that which is shown by each 

 day separately. Thus I have found that the range of the 

 mean lunar diurnal oscillation for the lunation December 

 16, 1858, to January 15, 1859, ^'^ Trevandrum, was i''25, 

 while the ranges of the mean oscillations for the quarter 

 lunations varied from \'"jo to 2'7o, these quarter luna- 

 tions giving exactly the same laws as have been deduced 

 from eleven years observations for the same lunar 

 epochs. 



In order to understand the value of these results we 

 must compare them with the ranges of the solar diurnal 

 oscillations for the same months ; those for December, 

 1858, and January, 1859, were 2' "20 and 2''24 respectively. 

 And as on some days the lunar diurnal variation has 

 amounted to nearly 5''o (which is equivalent to 12' in 

 England with the smaller directive force), it appears that 

 the lunar action is sometimes greater than the solar action 

 at the magnetic equator. 



As long as the lunar diurnal action was considered to 

 be of the minute character first discovered, it was always 

 possible for the supporters of the heat thesis to suspect 

 that some small unknown heat action was in question. 

 Such an idea is no longer possible. The lunar is some- 

 times greater than the solar diurnal action ; and the 

 former is dependent for its magnitude on the light and 

 heat vibrations dtie to the sun shining on the place of the 

 magnetic needle.^ 



If the solar light and heat vibrations can increase the 

 magnetic action, there can be no difficulty in believing 



' Mr. Willoughby Smith's experiments show that the light vibrations of 

 the ether in selenium diminish in a very marked manner the electrical resist- 

 ance of the crystal ; and it does not seem improbable that the increase of the 

 lunar magnetic oscillation in sunlight may be due to some similar action. 



that these vibrations may in their turn suffer some modi- 

 fication of intensity. It would be difficult to measure 

 small variations of the sun' s light with sufficient accuracy 

 as yet, though Mr. Willoughby Smith has suggested a 

 selenium photometer for this end ; we can, however, 

 measure the variations of temperature, and the fact that 

 the direct heating action of the moon is inappreciable is 

 no longer sufficient to disprove the results of Madler, 

 Kreil, Park Harrison, and Balfour Stewart. We have in 

 fact a mode of lunar action with which M. Faye was 

 unacquainted and could not take into account. The 

 whole basis of his argument is therefore destroyed. 



The view now given opens up a wide field of inquiry, 

 and cosmic meteorology appears under another aspect. 

 I hope to be able at another time to present other facts 

 which seem to relate to magnetical and meteorological 

 phenomena. John Allan Broun 



THE NUTRITION OF BROS ERA ROTUNDI- 

 FOLIA 



DURING the summer of 1877 I began an experiment, 

 the results of which were given in a paper read 

 before the Linnean Society, January 17, 1878. A number 

 of Drosera Plants were freely supplied with meat, while 

 another set were kept without animal food. At the end 

 of the season the two sets were compared in various ways 

 with the object of deciding whether or not carnivorous 

 plants profit by an animal diet. In the abstract of my 

 paper published in Nature (vol. xvii. p. 222), it may be 

 seen how numerous were the advantages gained by the 

 fed plants. 



The further results of the experiment are not without 

 interest. 



The plants on Avhich I worked were cultivated in six 

 soup-plates, and after all the flower stems had been cut, the 

 plants in three of the plates were removed from the moss 

 in which they grew, and were counted and weighed. The 

 plants in the other plates were left undisturbed with the 

 object of comparing the new plants which should spring 

 up from the winter buds of the two sets in the following 

 year. They were removed to the hothouse in the course 

 of the autumn, in order that they might rapidly send up 

 the next year's leaves. By the middle of January, 1878, 

 it became quite clear that far more leaves were springing 

 up from the winter buds of the plants which had been 

 fed than from the others. Both sets of plants were now 

 kept without food, and on April 3 they were removed 

 from the plates, and carefully counted, dried, and 

 weighed. The following numbers give the result of the 

 examination : — 



It will be seen that there is only comparatively a small 

 difference (18 per cent.) between the number of not-fed 

 and fed plants. Numerous minute offsets were found 

 among both sets, and were all counted as separate plants- 

 But, judging either by the total or average weights, no 

 doubt can be entertained of the great advantage gained 

 by the fed plants. It is a striking fact that, in spite of 

 the far larger yield of flower stalks, seeds, &c., produced 

 during the previous summer by the fed plants, they were 

 nevertheless enabled to lay by a far greater store of 

 reserve-material than their not-fed competitors. 



