As will be seen from the table, those months in which air circulation departs from the simple Antarctic type are in general the most rainy. August certainly seems to be an exception, but this may be partly explained by the fact that only three months of August, in which all storm systems were "Antarctic," occurred during the 25 years, and also that amongst cyclonic developments are included "tropical troughs" and "dips," which may not always have extended far enough south to appreciably affect Victoria as storm systems. Figures 72-74 show typical specimens of some of the systems above referred to:—
| Ordinary Antarctic depression | 2nd June, 1911. |
| Antarctic cyclone | 2nd September, 1911. |
| Tropical dip | 22nd June, 1912. |
Effect of Air Flow from Tropical Belt.
A very common feature of good winters is the formation of trough-like depressions lying north and south and connecting with the tropical low-pressure belt. These cause extensive rains, falling mostly in front of the trough and extending sometimes right across the continent. They occasionally end by producing a strongly cyclonic circulation over South Australia, New South Wales, or Victoria, or all three, and when this happens the tropical influence seems to come to an end with cessation of trough rains and the production only of those due to the cyclonic system itself. A tongue of high pressure usually now presses from the westward along the northern boundary of the cyclone destroying the tropical connexion, and this is followed by the lessening intensity of the cyclone itself as if its main supply of energy had been cut off. This cyclonic formation and later peculiarities of development are most frequently seen when tropical influences only are operative.
To show the closeness of the relations between the winter rainfalls over the southern and south-eastern interior of Australia and the formation of these troughs over Central Australia, the daily weather charts for all the years available have been consulted, and counts made of the number of days during the six winter months (April-September), when the barometer readings at Alice Springs reduced to sea-level have been below 30.0 inches. In some of the cases the depressions would be due to purely tropical intrusions of low pressure, but it was not thought worth while to attempt to eliminate these, more especially as they would favour the production of the trough-like depression above referred to. The table shows that 19 times out of 25 the rule holds that winter rains are above or below normal according as the number of days occupied by trough depressions giving readings below 30.0 inches and passing over Alice Springs is above or below normal. Then the eight selected dry years for which charts are available give, for the number of days thus occupied, a mean only two-thirds as large as that given by the twelve wet years. A typical wet chart is that of 31st May, 1909, in which the tropical trough connexion is a prominent feature, and dry chart that of 6th September, 1907, in which the trough is wanting. See Figures 70 and 71.
It is perhaps necessary to explain here that the term "trough" is used in a special sense. It is customary to use the word "col" when referring to the space separating the isobars defining two "highs." Analogy is drawn from land contour. Two mountain ridges may be joined by a saddle or col, or separated by a valley, along which perhaps a river flows. So in considering the connexion or separation of two barometric high pressure
