The most prominent featuie in the recoveries of 

 the units liberated on June 26 was the movement al- 

 most exclusively to the north and east of release sta- 

 tions. Ayers et al. (1958) noted a similar movement 

 of drift bottles released June 28 and 29, 1955, in this 

 area. Only near the Michigan shore did any substan- 

 tial numbers drift to the south. In July, when the major 

 number of these bottles were adrift, the percentage of 

 southwest winds was high (fig. 8). 



A major change in circulation from that shown 

 by the recoveries from the May 13 releases had oc- 

 cuned in the western area of the lake by June 26 or 

 shortly thereafter. After May 13 the drift was to the 

 south whereas subsequent to the June 26 releases the 

 drift was to the north (figs. 11-12). 



The drift of units released on August 11 was par- 

 ticularly complicated (figs. 11-12). The movement 

 in general was from west to east, apparently under 

 the influence of the pcevailing westerly winds of Aug- 

 ust and September (fig. 8), but recoveries from re- 

 leases at adjacent stations frequently were many miles 

 apart on the east shores. An unusual combinaticm of 

 wind stress and density distribution probably caused 

 the widespread recovery. Evidence of exceptional 

 conditions comes from the major fluctuati(xis in water 

 temperatures first detected by the M/V Cisco off Lud- 

 ington on August 9, 2 days before the drift bottles 

 were released. Sharp decreases in the water temper- 

 ature of the upper layers continued until the surface- 

 water temperature had dropped to 4. 8°C. on August 14, 

 2 miles offshcxe north of Ludington. The "normal" 

 surface-water temperature in the area at this time of 

 year is near 20. 0° C. Ayers &. aL (1958) also found 

 the cold water along the east shore as well as a major 

 north -south current in the area on synoptic surveys of 

 August 9 and 10. They attributed this condition to the 

 following situation: "When effective wind stress in the 

 Straits is from directions too far north or east of normal, 

 wind -driven surface water transport (current) is directed 

 back into Lake Michigan where it induces modifica- 

 tion of the normal current pattern. " 



The pronounced south -bound east -shore current 

 described by Ayers is not substantiated by returns of 

 drift bottles with drags released August 11 at the second 

 and fourth stations west of Ludington (fig. 11). Bot- 

 tles released at these stations were recovered mainly 

 to the northeast and had traveled at relatively rapid 

 rates. Ayers et al. (1958) believed, however, that 



shortly after August 10 (the date of his last synoptic 

 survey) the north-bound current was soon re-estab- 

 lished. This belief, if true, could account for the 

 recovery locations of bottles released at stations 2 

 and 4, but would not, however, account for recov- 

 ery to the south of bottles released at station 3. 



The recoveries from the bottles released in Sept- 

 tember 26 and November 8 exhibited a general west- 

 to-east drift caused presumably by the westerly winds 

 that predominated in October and November (fig. 8). 



Frankfort -Sturgeon Bay transect . — The transects 

 from Frankfort to Sturgeon Bay offer still another ex- 

 ample of the considerable variability of surface cur- 

 rents and also give another example of the similarity 

 of results from releases of bottles with drags and bot- 

 tles ballasted with sand. 



The recoveries of the releases made on the tran- 

 sect on May 16 and 17 (strong winds prevented com- 

 pletion of the transect on May 16) disclose a pattern 

 remarkably similar to the recoveries from the Luding- 

 ton -Manitowoc releases 3 days earlier. The littoral 

 current on the east shore was north -bound and that on 

 the west shore was south-bound. Recoveries from the 

 intermediate stations were made mostly to the east 

 and north, particularly for stations to the east of the 

 middle of the lake (figs. 13-14). Recoveries from the 

 releases of June 23 disclose that the surface drift had 

 shifted from south to north in the western area of the 

 lake. The distribution of recoveries for this transect 

 was similar to that of the recoveries from the Luding- 

 ton -Manitowoc transect 3 days later in that the drift 

 was almost entfrely northward. 



The bottles released on August 19 moved in a 

 more consistent pattern than did those of the releases 

 8 days earlier on the Ludington-Manitowoc transect 

 (figs. 13-14). Thermograph records from the M/V 

 Cisco on August 15 off Ludington show an increase of 

 surface-water temperatures over those during the major 

 upwelling a week earlier. By August 19 surface tem- 

 peratures off Frankfort (19° to 24° C. ) were again near 

 normal for mid -August and drift -bottle movement to 

 the north from releases near shore off Frankfort suggest 

 that the "normal" north -bound cunent had become re- 

 established along the east shore. 



The drift from the September 22 and November 4 

 liberations was generally from west to east except in 



24 



