An important question about this bifurcation arises: When does the periodic solution disappear? Is it again at the old bifurcation value of L=125’ or has the periodic harvesting changed this ? To examine this question, some students looked at longer and longer time scales (see Figure 5). Their results suggest that the bifurcation value occurs at a lower value than L = 125, but since we have no analytic methods available to deal with this situation (at least not in a sophomore-level course), it’s hard to tell. At the time the first author had to grade this, there was some dispute as to what actually was going on—the course decided on (always a useful technique ) was toraise doubt in the students’ minds, to encourage further exploration (and buy some time). Those students who had asserted that the bifurcation value for (2) was still L=125 were shown Figure5, right; those who asserted that the bifurcation must have occurred earlier were encouraged to think about computer round-off as well as the inherent error in Euler’s method (or any numerical technique).
This parallels the real-world situation: if fishing is allowed at exactly the bifurcation value, the population lives on the brink of disaster. If a sudden fluctuation in the population causes the population to rise a bit (a fisherman takes a day off, a shark dies or gets indigestion, twins are born!),the effect will be flattened out by the downward tendency of all solutions above the unique periodic solution. But any fluctuation (illegal over-harvesting, a sudden epidemic, poor breeding conditions, etc.) that pushes the population below the periodic solution will result in extinction. Similarly ,this is how computer error has an effect: an overestimate of the population will eventually disappear, but an underestimate will get magnified immediately.