BIED MIGEATION". 29 



imperfect and frequently at fault . Doubtless a similar but vastly more 

 acute sense enables the murres, flying from home and circling wide 

 over the water, to keep m mind the direction of then* nests and return 

 to them without the aid of sight. 



But even the birds' sense of direction is not inf aUible. Reports from 

 hghthouses in southern Florida show that bhds leave Cuba on cloudy 

 nights, when they can not possibly see the Florida shores, and safely 

 reach then* destmation, provided no change occurs in the weather. 

 But at fickle equinoctial time many flocks starting out mider 

 auspicious skies find themselves suddenly caught by a tempest. 

 Buffeted by the wind and their sense of direction lost, these birds fall 

 easy victims to the lure of the Hghthouse. Many are killed by the 

 impact, but many more settle on the framework or foundation until 

 the storm ceases or the coming of daylight allows them to recover 

 theic bearings. 



A favorite theory of many American ornithologists is that coast 

 lines, momitain chams, and especially the courses of the larger rivers 

 and theh tributaries form well-marked highways along which birds 

 return to previous nesting sites. According to this theory, a bird 

 breedmg in northern Indiana would in its fall migration pass down 

 the nearest httle rivulet or creek to the Wabash River, thence to the 

 Ohio, and reaching the Mississippi would follow its course to the Gulf 

 of Mexico, and would use the same route reversed for the return trip 

 in the spring. The fact is that each comity in the Central States con- 

 tains nesting birds which at the beginning of the fall migration scatter 

 toward hah the points of the compass; indeed, it would be safe to say 

 aU the points of the compass, as some young herons preface their 

 regular journey south with a Httle pleasure trip to the unexplored 

 north. In fall most of the migrant land birds breedmg in New 

 England move southwest in a line approximately parallel with the 

 Allegheny Mountains, but we can not argue from this that the route 

 is selected so that mountains will serve as a guide, because at this 

 very time thousands of birds reared ui Indiana, lUinois, and to the 

 northwestward are crossing these mountains at right angles to visit 

 South CaroHna and Georgia. This is shown specifically in the case 

 of the palm warl^lers. They winter in the Gulf States from Louisiana 

 eastward and throughout the Greater Antilles to Porto Rico; they 

 nest ill Canada from the Mackenzie VaUcy to Newfoundland. To 

 migrate according to the "lay of the land," the Louisiana palm 

 warblers should follow up the broad open highway of the Mississi])pi 

 River to its sourc(5 and go thence to their brooding gromids, while the 

 warblers of the Antilles shoidd use the Alloglieny Mountains us a 

 guide. As a matter of \'iicA , tlio Louisiana bhds nest in Labrador and 

 thoH(; from tht; Aniillos cut diagotniUy across the United States to 

 summer in central Canada. Those two routes of pahn warblers 



