is clearly evident as indicated by the depressed isotherms of the temperature 

 sections in figure 1. The horizontal dimensions across the prime eddy are 400 

 to 600 km with variations occurring during a single monsoon season (Bruce, 

 1970) as well as yearly differences which will be described here. 



An eddy of smaller horizontal dimensions which appears to be associated 

 with the prime eddy was found each year (1979-1979) off Socotra between ]2^H 

 to 150N. During some years such as 1976 and 1979 falso 1970 (Bruce, 1973)] an 

 eddy was observed south of about S^N and adjacent to the southern boundary of 

 the prime eddy and to the east African coast. 



Observations during several surveys within the last fifteen years indicate 

 that the northeastward current flowing alongshore (Somali Current) is clearly 

 part of the eddy field found each southwest monsoon off the Somali coast 

 (Swallow and Bruce, 1966; Bruce, 1968, 1973). The current diverges from the 

 coast turning eastward about 9°li to 10°N each year and during some years also 

 at 30N to 5°N forming a southern eddy. Farther offshore (550E to 580E), it 

 turns southward and then back toward the shore. All past measurements known 

 to the author indicate that during the southwest monsoon, a clockwise "warm" 

 eddy (prime eddy) of this general description occurs within the Somali Basin. 

 Pronounced upwelling with surface temperatures as low as 13°C (Warren et_ al^. , 

 1965) is found in the region where the strong coastal current turns offshore. 



The program has utilized ships of opportunity (EXXON tankers) en route 

 along the sea lane between the Persian Gulf and South Africa. These obtained 

 a temperature-depth (0-450 m) section with expendable bathythermograph probes 

 (XBTs) on an average of approximately every three weeks along essentially the 

 same track. The track location is extremely fortuitous in that it passes 

 directly through the central region of the eddy field and thus provides an 

 excellent means of monitoring the growth and decay of the eddies formed during 

 each southwest monsoon. The measurements were obtained by special observers 

 who were placed aboard at Cape Town, South Africa, and made a round trip to 

 the Persian Gulf, thus obtaining two sections per trip. Altogether 55 sections 

 were completed (figure 2). The closely spaced stations (20-30 km apart) 

 necessary to observe the small scale features of the temperature structure 

 essentially require a full time observer who was also needed to maintain good 

 quality control of the data and record, at each station, wind velocity, ship's 

 set by currents, surface salinity samples, etc. 



With each of the sections shown in figure 1 (which represents the fully 

 developed southwest monsoon eddy system) is given a schematic representation 

 of the circulation pattern of the near-surface water (upper 150 m). This 

 estimate is aided by previous surveys in this area (Bruce, 1968, 1970) during 

 which time the structure of the eddies to the east and west of the tanker 

 track was observed. The complete time sequence showing the changes occurring 

 in the thermal structure of the eddy field is given in figure 2. In late 

 March and April no large scale horizontal gradients generally are evident in 

 the upper thermocline, whereas in the late June - early July sections the 

 prime eddy is clearly discernible roughly between 5°H to 10°N. Then during 

 July and August (the periods of maximum wind strength) the development inten- 

 sifies with a deepening of the mixed layer in the central regions of the prime 

 eddy (centered approximately 8ON) and the Socotra eddy (12°N to 14°N). By 



