122 DISCOVERY REPORTS 



of observations between every two degrees of latitude: the last separately for each 

 station as the residual of the observations available for a fortnight prior to our arrival 

 off the coastal strip in question. 



Reduction of wind forces to residuals followed the normal practice of vector averaging : 

 the directional velocities expressed as miles per hour were first resolved along the 

 cardinal points of the compass; these were summed and reconverted into an aggregate 

 resultant which was then divided by the original number of observations. 



This method of reducing data gives a working idea of wind conditions, even though 

 the number of observations represented by each arrow in Fig. 4 may vary according 

 to the length of time spent by the ship off each stretch of coast. In certain instances, 

 when a rapid change in wind was followed within a few hours by changes in hydrological 

 conditions, the data before and after the change have been plotted separately. This has 

 been done off Antofagasta and off northern Peru where the ship covered the same ground 

 two or three times, and the data on the voyage northward in May to July have been plotted 

 independently of the voyage southward in August to September. 



Specific instances of the effect of wind upon the ship's drift and upon upwelling are 

 discussed later ; here it may be well to take a general view of the winds prevailing on the 

 west coast at the time of our visit. 



Fig. 4 will show the existence of natural regions in each of which the majority of 

 winds usually blow in a set direction. We may differentiate between the following: 

 south of lat. 38° S winds are irregular in force and direction in the months of both May 

 and September. This region lies in the circumpolar tract of the "westerlies", known 

 also as the "roaring forties". North of 38° S winds are more regular. They appear to 

 have least force over land and most far out to sea. The weakest winds over the coast are 

 naturally least certain in direction, but those immediately offshore blow almost without 

 exception parallel to the coast, i.e. they are southerly winds off Chile and south-easterly 

 off Peru. 



Farther from the Chilean coast, in August at any rate, the winds showed a departure 

 from the coast-line direction : they veered from southerly inshore to south-easterly as the 

 distance from land increased. These winds between the parallels of 32 and 4° S lie in 

 the path of the south-east trades, and it would appear that while the south-east winds 

 off Peru and far out to sea off Chile may be classed as normal trades, the coastal winds 

 represent a departure from them. Such a distribution of winds is normal for the circula- 

 tion in the eastern South Pacific; Brooks (1929) thus summarizes the position as it is 

 understood at the present day: 



The centre of the South Pacific anticyclone lies off the coast in the latitude of 30° S, and pressure 

 is highest in this latitude, decreasing slowly northward and more rapidly southward. The northern 

 part of the region as far as 30° S is throughout the year under the influence of the south-east trade 

 winds, which cross the equator and blow into the Gulf of Panama as south-westerly winds. South 

 of 30° the winds are south-westerly, rapidly becoming westerly and increasing greatly in force as the 

 "roaring forties" are approached.. . . 



From the belt of high pressure in about 30° S pressure decreases steadily southward, very rapidly 

 between about 40° and 50°. Here strong westerly winds prevail throughout the year, often of gale 



