This might be more fun if you presented a couple of challanges to your readers that you don't provide the solution for - in other words provide the raw data, a picture of your solution and the page reference in Tufte and see what your readers come up with.
;~)

Nice! I have just a couple of criticisms:

I've tried using special symbols before, and they're very dangerous, because they aren't stable when imported into other programs. You're liable to find that your tick marks move if you put the chart into a Powerpoint presentation, either grossly or (what is even more dangerous) subtly. My approach to the tick marks in the dot dash plot is to use Excel's error bars instead. These plot precisely, and can be made any colour, and hair thin if desired.

My second criticism is that you've misunderstood the role of the "air gap" in Tufte's minimalist quartile plot. It's not just for show, but encodes the upper and lower quartile (the ends of the whiskers are the minimum and maximum). So instead of a custom marker with white space around, you really need three markers: median, upper, and lower quartile, with the median marker being one of Excel's standard types (see my worries about wandering Autoshapes above). The upper and lower quartiles should then be invisible, but have error bars reaching up and down to the minimum and maximum.

However, I disagree with Professor T (gasp!) and always make my interquartile range bold, as that's what I want my viewers to focus on. The minimalist quartile plot does shows all the information for those prepared to pay attention, but the "air gap" places the two inner interquartile ranges on a lower perceptual "layer" than the two outer ranges, when the opposite should be the case.

I agree with Derek on the emphasis owed to the interquartile range. But the obsession to reduce ink must be very strong. On the facing page in 'Visual Display of Quantitative Information' (p. 124) ET demonstrates various ways to display the quartiles. My favorite is the one with bolder lines for the interquartile, but I would have added a discrete marker (a small cross perhaps) for the median. Leaving a tiny gap has the same effect as offsetting the line slightly to denote the IQ range: I squint to see if the graphic was rasterized, and the gap or offset was caused, for example, by imprecise photocopying. If you erase too much, you risk losing too much information. As the man says, it's about "reasonable" data-ink maximizing.

Jon Peltier, Microsoft Excel MVP
http://PeltierTech.com

The gray bar chart with white gridlines in front of the bars has always struck me as one of those optical illusions, and I watch the chart to see what's going to slowly change. I lose focus on what's being charted. Too little of a good thing (ink).

Derek C wrote >> You’re liable to find that your tick marks move if you put the chart into a Powerpoint presentation, either grossly or (what is even more dangerous) subtly.

I use Excel a lot as a drawing tool. It works best of all the Office tools, because there's a clear distinction between the working plane and the drawing plane. When "importing" any Excel results (graphs, drawings or even bits of spreadsheets) into Powerpoint, I always use Paste Special/Bitmap. This prevents the distortion that Dereks speaks of - and has the added advantage of preventing my Excel logic being inspected (and manipulated) by the PowerPoint user!

I came across this site as a link through from the Edward Tufte site. As I offered on the Ask E.T. bulletin board, if anyone is interested in the Excel version of the charts you've shown here, I'd be more than happy to send them through. Great idea, by the way, on the use of the REPT function - I'll be using that a lot more!

I'm glad I stumbled into this site today. I think my excel charts will look much better from here on in. There are some great ideas here.

For a minimalist quartile plot that highlights the inner quartile, try using solid lines for the inner quartile error bars and dashed lines for the outer quartile error bars. (If you need painstaking precision, you'll may want to put in a 2pt dash marker on the upper and lower quartile datapoints so that you can tell where the line ends and the dashed line begins.)

I had first tried making the inner quartile a bolder line, but didn't like how it thickened the line to one side only. The dash-to-solid line is much less distracting.