There are many factors that can affect peak flow or time to peak  beyond watershed shape.  If the channel gradient and stream type are the same, the elongated watershed is apt to have an extented time for water to,travel further, so peak may be somewhat extended, and perhaps not as high as the subwatersheds and tributaries enter over a longer distance.  But from experience, elongated drainages due to gullying are apt to have low sinuosity and no access to floodplain, so their stream velocity, time to peak may actually be faster and as well higher peak as just an example.  If you are familiar with the unit hydrograph training, it has some of the basic elements in discussing effective rainfall and estimating time to peak and time of travel concepts for individual stream or drainage segments.  Watershed length and channel length are quite different factors, so trying to generalize based on basin form metrics helps to understand concepts, but in real circumstances, can lead to errors.  Two watersheds with same relief, but one with a uniform channel gradient will respond quite differently from one with abrupt falls or cascades and otherwise much lower gradient or braided channels.  

I think Schumm might be a good author to address some of these elements, as well as Horton, Strahler and Shrieve on stream ordering and channel branching subject.

I agree with the answers of both Dr. William and Dr. Kevin. Considering all the other parameters being the same, it is quite a general rule that the occurrence of  flash floods is more in circular watersheds and there is extended time for the peak discharge in the elongated watersheds. There are class books written by the greats like Horton, Strahler and Schumm.

thanks Dr. William ,Dr. Kevin and N.B.Narsimha Prasadji for your valuable answers.