485 



Beciufort's scale, was also recorded every two Inmrs with each observation of direction. The record of 

 mean wind velocity, in miles per hour, deduced from 2-hourly readings of the cup anemometer, is 

 practically complete from April, 1902, till the end of the vessel's stay in Winter Quarters, the breaks in 

 its continuity amounting altogether to only a few days, during which the instrument was disabled or 

 under repair. The series of observations of velocity by the "sight-indicating" pressure tube is somewhat 

 less complete, whilst the records of the recording pressure-tube anemometer exhibit frequent lacunae, which 

 in many instances extend over days and in several others over weeks. 



DIRECTION OF THE WIND. 



The observations of wind direction, irrespective of force, were first dealt with. 



All the observations for each month were first of all grouped under thirty-two points, and the percentage 

 of their frequency of occurrence under each point determined. It was at once seen that this was a greater 

 degree of refinement than the observations themselves would warrant, and it was evident from the 

 infrequency with which observations were recorded under the intermediate points that the observers had 

 practically limited themselves to sixteen points in making their observations. The observations under 

 each intermediate point were therefore allocated to the points on either side in proportion to the number 

 of observations already credited to them. Thus if five observations of E. by N. winds were to be 

 distributed between HN.E. and E., and there were already entered to E.X.E. ten observations and to 

 E. fifteen, then two of the five would have been given to E.N.E. and three to E., raising the number of 

 observations under the first-named point to twelve and that under the latter to eighteen. 



The results of this grouping are shown in Table I. 



TABLE I. Distribution of Wind Direction under Sixteen Points in Percentages. 



