Sorry to be a bit late with the answer. The woman screamed. I saw that the gator had a white bird in its maw in the bright sun. As I moved to sun angle I dialed in -1 stop. Why? The gator, in the middle of the frame and the metering pattern, was very dark and the dirt was pretty dark too. The 40D needs less light to yield the same histogram as the MIII so I instinctively put in some extra minus. In the first two frames of the four or five in the sequence in the sun were made at 1/4000 sec (all at f/5.6). In those the histogram was dead solid perfect. There was no room to the right. In the next three frames more of the dark gator filled the frame so you would expect and the exposure dropped to 1/3200 at f.5.6. One would expect that the whites would have been burned but instead, there was a bit more room on the highlight end of the histogram. Why? I think because of the angle of the sun. Even though the gator was in the sun the angle of the sun was reduced as it got closer to me and thus there was effectively less light on the scene. So far so good.
Now the animal slips into the dark shadow of the boardwalk. I change the -1 stop to -1/3 stop. Why? Not as much need to save the whites with the Cattle Egret now in the shade... There were about 20 burned pixels on the Cattle Egret that went away when I underexposed .2 stops during the BreezeBrowser conversion. Minus .1 probably would have been fine.
To my mind, I had nailed both exposures and this was confirmed by both the histogram and the very few flashing highlights in the shadow images. I showed the photos to a guy and to the woman who had screamed. I explained my exposure choices to them. I saw the guy the next day and he confided in me that when I had walked away the woman had said to him, "What the **** does he know?" I thought that that was pretty damned funny.
So my answers would be: 40D, -1, and -1/3. Had I had a MIII in my hand, they would have been: MIII, -2/3, zero.
If I had had a D-300, it would have been D-300, -1 1/3, and -2/3 (though I have never made an image with this camera I am a good listener...)
If you used less light for the images in the shade than for the images in the sun, you lose 20 points. The same goes if you added light at any time. If you kept the same exposure, subtract 10 points. Lots of folks did pretty well and got the concept that they needed to subtract more light in the sun. And for Mr. Curry-check, the exposure for the shadowed images was 1/200 sec. at f.5.6. That's four full stops (4 1/3 from the first images) lighter than in the sun. So either 12 or 13 clicks would have been needed. Sorry Charlie; no time for that!
The good news is that the original ABP will eventually be reprinted in China though we have run into a dozen delaying problems; it contains the finest treatment of exposure theory on the planet.
Additional questions or comments welcome.
ps: to Christian: yes, as Blake said, Evaluative metering. And anyone using spot-metering in this situation would be worse off than the manual exposure folks, AND, I would bet anything that anyone using a spot or an incident meter could not figure the right exposures in these two situations even if they had an hour to get it right!