A migration route, therefore, may be anything from a narrow path closely adhering to some definite geographical feature, such as a river valley or a coastline, to a broad boulevard that leads in the desired direction and follows only the general trend of the land mass. Oceanic routes appear to be special cases that are not fully understood at the present time. Also it must be remembered that all the main routes contain a multitude of tributary and separate minor routes. In fact, with the entire continent of North America crossed by migratory birds, the different groups or species frequently follow lines that may repeatedly intersect those taken by others of their own kind or by other species. The arterial or trunk routes, therefore, must be considered merely as indicating paths of migration on which the tendency to concentrate is particularly noticeable. =PATTERNS OF MIGRATION= Band recoveries, netting records, and personal observations help us to critically examine migration routes and probe deeper into the origin and evolution of these pathways. We are beginning to realize certain deviations occur from the "normal" north and south movements expected in most species. In the previous section on routes, we touched briefly on the fact that some routes are not poleward at all, but in some other direction. We know that many migrants do not stop at the exact localities year after year but they probably do follow the same general course each season. After many years of observations a pattern emerges for that population, species, or group of species. In this section we would like to take a closer look at some of the interesting patterns (or "eccentric routes" as Cook (1915a) referred to them) in migration that birds are annually to travel from breeding to wintering grounds and back again. In many cases, the causative agents are unknown or pure conjecture, but in others, sound biological principles can be put forth that may indicate why a particular species could have evolved the specific pattern it exhibits. Loops Many species do not return north in the spring over the same route they used in the fall; rather, they fly around an enormous loop or ellipse. Cook (1915a) considered food as the primary factor in determining the course birds took between winter and summer ranges. Individuals that returned by the same route and did not find sufficient food for their needs at that time were eliminated from the population, and only progeny from individuals that took a different course with sufficient food lived to build the tradition of a loop migration. Other investigators consider prevailing winds a major factor in the evolution of loop migration. Whatever the reason may be, it has most likely evolved separately in each species to satisfy its particular needs, and the fact that this pattern occurs all over the world in completely unrelated species is a good illustration of convergent evolution.