The observations on which these conclusions are founded include sixty-four years of the present century: too short a period on which to build a theory; but as no records exist earlier than 1810, it is by future observation only that the conclusions can be tested. Meanwhile meteorological observers will be watchful, especially of the rainfall, for India is a country which affords singularly favourable opportunities for a comprehensive system of observations. The other paper referred to above is by Mr H. F. Blanford, Meteorological Reporter to the government of India. He points out that Dr Hunter’s views apply exclusively to Southern India, and that in Northern India famines are most frequent at the epochs of most sun-spots. This lack of agreement between two competent authorities shews how great is the need for a lengthened series of observations. In the recent Arctic Expedition twenty-five species of fossil plants were discovered in Grinnell Land by Captain Feilden. They are of the period described by geologists as Miocene, and can be identified with species of the same period found in Europe, in North-western America, and in Asia. Among them are two kinds of _Equisetum_, poplar, birch, elm, and pine. It was suggested at a meeting of the Geological Society that the bed of lignite in which these remains were met with was in remote ages a large peat-moss, probably containing a lake in which the water-lilies grew, while on its muddy shores the large reeds and sedges and birches and poplars flourished. The drier spots and neighbouring chains of hills were probably occupied by the pines and firs, associated with elm and hazel. Among all these which indicate a primeval forest, the only sign of animal life discovered was a solitary wing-case of a beetle. When water-lilies were growing in that now desolate region, fresh-water must have filled the ponds and lakes. Captain Feilden’s discovery may be taken as additional evidence of a change of climate, which the palæontologists and physicists who are now discussing that interesting question will not fail to make use of on fitting occasion.