T think it would be premature to issue forecasts at present 

 until we get to more tlioroughly understand the behaviour of 

 storms between Leeuwin and Eucla, and the only way to do 

 this is by careful inspection every day of isobar maps of the 

 entire continent, such as are displayed in the hall of the Post- 

 office, and by carefully noting the weather accompanying cer- 

 tain recognised forms of disturbance. Melbourne is much 

 more favourably situated than we for forecasting, since, as I 

 I have pointed out, storms are to a great extent regular in their 

 passage after passing Eucla, and thus observers at Melbourne 

 have very often a good thirty-six hours' warning at least. Of 

 course a number of circumstances may easily tend to render a 

 forecast at Melbourne incorrect — such, for instance, as the 

 case of a storm-centre off the south of Tasmania moving up or 

 expanding northwards. This has happened on several occa- 

 sions, and of course no warning of this could possibly be 

 obtained. The cause of such a movement would probably be 

 found far away in the Southern Ocean. 



As a matter of fact, forecasts are published regularly in the 

 Argus by the Victorian Government Astronomer, and, I believe, 

 are to a very great extent trustworthy. With our present 

 knowledge of the general laws of storms, and the behaviour of 

 certain particular types of Australian cyclones, I believe it 

 would be possible to issue forecasts for twenty-four hours 

 ahead, of which a large percentage would be correct. Owing 

 to the difficulties I have pointed out, however, the forecasts 

 could not be absolutely depended on ; but we may look for- 

 ward confidently to some time in the near future when fresh 

 discoveries and more elaborate research will enable the prac- 

 tical meteorologist to foretell with comparative certainty the 

 coming climatic conditions for one, two, or three days, if not 

 for weeks. 



