the western part of the lake (figs. 13-14). The mass 

 landing on the Wisconsin shore from the 2 westernmost 

 stations from the September 22 releases may have been 

 caused by east winds. The U. S. Coast Guard station 

 at Sturgeon Bay reported east winds from 2 days before 

 until 1 day after release. 



If Ayers et al. (1958) were correct in their con- 

 clusion that winds too far north or east of normal in 

 the Straits of Mackinac cause modification of the nor- 

 mal flow (normal being north-bound off the east shore), 

 north or east winds must have been strong on or about 

 September 22 to cause the prominent southward flow 

 off Frankfort (figs. 13-14). Wind observations at U.S. 

 Coast Guard stations on Beaver Island and at Manistee, 

 Michigan (table 6), support their conclusion. 



Charlevoix -Manistique transect . —The returns 

 from releases of bottles along this, the most northerly 

 of transects, exhibited many similarities to those from 

 releases at corresponding dates along more southerly 

 transects. The data for bottles with drags and for sand- 

 ballasted bottles agreed well again except for inshore 

 stations off Charlevoix. 



The recoveries of bottles released on June 5 re- 

 sembled those from the first Ludington -Manitowoc and 

 Frankfort -Sturgeon Bay releases in that the drift was 

 essentially counterclockwise (figs. 15-16). The direc- 

 tion of drift of sand -ballasted bottles and bottles with 

 drags differed at the first two stations off Charlevoix on 

 June 5 (figs. 15-16). The bottles with drags drifted to 

 the northeast whereas the sand -ballasted bottles were 

 recovered mostly from the southwest. 



On July 14, nearly 6 weeks after the first releases 

 on this transect, the littoral cmrent on the west shcxe 

 was to the northeast instead of southwest. The returns 

 from the remaining stations were similar to those of 

 the June 5 transect (figs. 15-16). 



The recoveries from releases made on the Sep- 

 tember 5 transect exhibit few clear trends (figs. 17-18). 

 No strong Unoral cunents were evident, and many bot- 

 tles released in the center of the lake traveled to the 

 west shore and north to the islands. 



West -to-east drift had reappeared by late Octo- 

 ber and November, as evidenced by recoveries of the 

 October 19 releases (figs. 17-18). 



Rate of drift. --The average drift rate of the bot- 



tles with drags was 1. 5 mUes per day as compared 

 to 1. 8 miles per day for the sand -ballasted bottles 

 (table 2). On 3 of 16transects, however, the drift 

 rate of the bottles with drags was higher than that 

 of the sand-ballasted bottles. A very high rate 

 of drift for a bottle with drag was for one that travel- 

 ed 64 miles across the lake from west to east in 10 

 days, a drift rate of 6.4 miles per day (table 17). 

 A sand -ballasted bottle released at an adjacent sta- 

 tion made a 65 -mile crossing in 9 days, a drift rate 

 of 7. 2 miles per day (table 18). These bottles were 

 released on November 8 at stations 7 and 8 respec- 

 tively on the Ludington -Manitowoc transect. The 

 Sturgeon Bay Coast Guard station reported prevail- 

 ing westerly winds for the 10 days following release 

 of these bottles on November 8, and southwest winds 

 of Beaufort 10 (55 to 63 mUes per hour) on November 

 16. The maximum drift rate for sand -ballasted bot- 

 tles (11.0 miles per day) occurred for 2 bottles re- 

 leased on November 4, approximately 15 mUes east 

 of Sturgeon Bay, Wisconsin (table 20), Both types of 

 units released at station 4(approximately 15 mUes 

 west of Ludington, Michigan) on June 26 traveled to 

 the north at high rates of drift; a sand -ballasted bot- 

 tle traveled 115 miles in 14 days (8. 2 miles per day, 

 table 18) and a bottle with drag 87 miles in 14 days 

 (6. 2 mUes per day, table 17). During these 14 days 

 the Manistee Coast Guard station reported 57 obser- 

 vations of winds with a southerly component (wind 

 from the south), 11 northerly, 26 westerly, and 16 

 easterly; the south winds were slightly stronger than 

 those from any other direction. The overwhelmingly 

 predominance of south winds over north winds and 

 the relatively greater number of west winds over east 

 winds doubtless account for this rapid drift rate to the 

 northeast. 



General characteristics. --The north -bound lit- 

 toral cunent found off Grand Haven in 1954 was evi- 

 dent in 1955 as is shown by the recoveries from bot- 

 Ues released April 26 (figs, 9-10), 



In central areas of the lake, drift was generally 

 west to east and south to north. Greatest deviation 

 from this pattern followed the release of bottles Aug- 

 ust 11 on the Ludington-Manitowoc crossing (figs, 11- 

 12) when drift to the south was equally as prominent 

 as drift to the north. 



In the northern areas of the lake, drift was 

 mainly west to east. 



27 



