H.—34

weather occupies a somewhat unique position. Volcanic eruptions, geological changes, and other phenomena undoubtedly influence the weather of the world, but for day-to-day practical purposes it may be said to be governed by the existing state of the atmosphere and the earth's surface and the influence thereon of solar radiation. There are certain known slight fluctuations in solar radiation, and in past ages there may have been considerable changes. But, again, as regards daily weather-variations these may be, to a large extent at least, neglected. Seasonal changes are very important, but they are regular and can be allowed for. The weather of to-morrow may thus be said to be implicit in that of to-day. The ideal attainment would therefore be to be able to calculate it with precision by the application of known physical laws. But the atmosphere is so complex and extensive a body that hitherto it has been possible to obtain only partial solutions to its equations of motion, and never yet have we had complete knowledge of its state. From the oceans and the polar regions, particularly, the information is always scanty. And there is no doubt that conditions out almost to the extreme limits of the atmosphere are important. It has thus not been possible to put forecasting on an exact basis. The existing systems depend on the securing of as complete a knowledge as feasible of the state of the weather over an area extending as far possible in all directions around the region for which the forecast is required, and then deducing the changes that will occur. The latter is done principally on the basis of general physical principles and a knowledge of the processes of the atmosphere but, also, unfortunately, for the reasons stated above, partly on empirical grounds. The forecaster has to use his knowledge of previous situations which most nearly resembled the existing one in estimating the coming changes. The meterologist is fully aware that the result falls short of perfection, and if circumstances permitted he would prefer to defer the issuing of regular public forecasts until the system had been placed on a sounder basis. Forecasts are, however, sufficiently accurate to create an insistent public demand for them, and the only thing to do is to endeavour continuously to improve their accuracy, and this is being done. The principal advances naturally come from Europe, where science and civilization are most highly organized. The advances have all been in the way of collecting more and more information at more frequent intervals and the application of the methods of physical research to the information thus gathered. Weather phenomena are due principally to the inter-play of air masses of different origin and consequently of different characteristics as regards temperature and moisture content. The Norwegian School of Meteorologists, particularly, has become extremely expert at deducing the nature of the interaction of these currents and also the changes which will take place in the air-masses themselves as they pass over ocean or land surfaces or meet with the obstructions caused by mountain-ranges or plateaux. Their forecasts for short periods of from an hour or two to about twelve hours are usually very accurate, and by dividing the forecast area into a large number of small districts they can be made very precise. For the longer periods the limitations mentioned earlier become increasingly important and the accuracy begins to fall off rapidly. For the value of these developments to be passed on to the public it is necessary that the collection and dissemination of information should be both very complete and rapid. This is secured largely by the use of wireless telegraphy. Four weather charts are prepared daily, two of them covering a large part of the Northern Hemisphere. Here we must mention one of the unavoidable limitations of weather forecasts, and that is the ililliculty of describing the weather. On most days there is so great a variation from place to place that merely to describe the weather over a country like New Zealand at any one instant would require a large number of words, while to give a detailed account of that occurring in twenty-four hours would require pages. This is a difficulty that is seldom realized, especially by city folk. When showers are moving at intervals across the country for example, it never will be possible in a general forecast to tell the people in all the suburbs of a large city at precisely what times to expect rain. It will be clear from the foregoing that the success of a forecasting-service depends first and foremost on the availability of free and rapid means of intercommunication by telegraph, cable, and wireless. This is generally realized. In Australia, for example, the telegraph and .wireless services are available to the Meteorological Bureau free of charge, the only limitation being that the use should be reasonable. The amount of traffic thus handled is between fifteen and twenty times that in New Zealand. The leading wireless - telegraph companies throughout the world have agreed to transmit weather messages at half the ordinary rates and priority is accorded to them. The private cable companies in the China seas transmit weather cables free of charge. In England quarter-rates are charged for meteorological messages. In New Zealand, on the other hand, there have in recent years been progressive restrictions. The Meteorological Office is charged, in accordance with the present system, for all services by the Post and Telegraph Department, which has a monopoly of communications. The item for telegrams is consequently the largest in the budget of the Branch and the first to be attacked when an economy campaign is in progress. Some years ago morning reports from all over the Dominion used to be displayed daily for the public information in all the important ports. It was thus possible for shipmasters and others to form a reasonably complete view of the state of the weather over the country and interpret intelligently the official forecast. The number of places to which the report is not sent has been greatly reduced, and not more than twelve stations may be included in the list of reports to any one place. Such a service is quite inadequate. A more complete one would be of great use to motorists and aviators, as well as other interests. Again, the number of towns to which the mid-day forecast is sent has been progressively restricted. The official forecast itself has to be sent in a code such as is used in no other country and which is quite unsuitable. It is true that the wireless-broadcasting services have become available for disseminating information, and in many respects they have advantages over line services or issues in Morse code. But a telephoned broadcast is not a good medium for transmitting actual weather reports, and in general does not altogether replace the written word. As regards information received by the Meteorological Office, there has also been a considerable reduction in the number of reporting-stations.

69