the graph isn't really inversed. hyperpolarization occurs just before you get another depolarization of the membrane. and the depolarizations jump from node to node (Nodes of Ranvier).
the picture is kind of goofy, no offense to TBR, but their diagrams aren't exactly the best to look at (and neither are EKs by any means). i mean they aren't textbook quality illustrations, but they get the point across.
the main idea to take home is that these actions potentials jump from node to node. in between the nodes, there are no membrane channels and the membranes are insulated to prevent leakage. this makes propagation (the movement) of the signal from one area to the other more efficient. so as a sign originates from the axon hillock it moves down to the end of the axon, to the synapse, like the illustration.
lets say you had
Axon hillock -> X -> Y -> Z -> Synapse/End of axon
At time t=0 there would be depolarization ONLY at the axon hillock
At time t=1 there would be depolarization ONLY at the X
t = 2 ... only at Y
t = 3 ... only at Z
...
maybe i went too in depth based on your question. but you have to remember that the graph in 1-9 is a continuum, like a clock. on a clock, after 24 hours (or if you have an analog clock 12 hours), the cycle repeats itself. what happens after A through E in figure 1-9? well, you can get another action potential so you can go from A through E again (on that figure time 0 msec to 8 msec). in conduction of the action potential in an axon, you go from A-E at the axon hillock, then A-E at X, then A-E at Y, and so on.
i will admit that the second action potential in 1-10 looks a little weird since normally hyperpolarization can't cause an action potential. and rather depolarization above the threshold can only cause an action potential.