Yes, I get that... But it wasn't doing anything in the simulation of the Uno or Leonardo circuits. When DTS went low, reset simply followed DTS. It didn't go low, and then high again.
I haven't figured out the problem with the Leonardo schematic, and I've given up on it, but I have finally had some success simulating the Uno schematic.
Here is the Uno schematic:
And here is the simulation of it:
Sim: Circuit Simulator Applet
As you can see it now peaks slightly above 5V. So far so good.
But look at what happens when I set Vcc to 3.3V:
Sim: Circuit Simulator Applet
That is with a diode forward voltage of 700mv. Approximately the fv of the diode Crossroads says is used on the Uno:
http://forum.arduino.cc/index.php?topic=86389.msg1178112#msg1178112
In order to keep the spikes below 3.6v, I have to set the diode's fv to less than 325mv.
I suppose that doesn't matter on the Uno (or the boards that use the 32U4) because the reset pin is special and can handle up to 13v regardless of Vcc. But I thought I'd mention it because I'm looking at using an Xmega and the reset pins on those are the same as any other pin, aside from a 25K internal pullup so their reset pin can't stand more than 3.8v max.
Anyway, I guess my problem is resolved.