Starting with your cylinders around arbitrary minor axes, I can see that your overall control in drawing the ellipses improves a great deal, and you're quite diligent in checking the alignment of those minor axes. While you do still have some trouble aligning to the given minor axis, you seem to still be able to align the two ellipses together in order to keep a single cohesive cylinder. Aligning to the first mark you'd drawn certainly still is important, as it allows us to plan out our cylinders ahead of time, but this is definitely good progress.

One thing I did notice however is that while [the homework section]() states in bold that you should vary the rate of foreshortening applied to your cylinders, you do this very minimally. Aside from the odd case, the quantity of which I can probably count on one hand, the page with 116-120 appears to be the only one that doesn't feature cylinders that pretty much have no foreshortening at all.

This is technically incorrect - foreshortening is represented by the shift both in the scale of one ellipse compared to the other, and in the degree. By looking at the rate at which the two of these change along the length of the cylinder, the viewer is able to estimate how long that cylinder is, in rough terms. If however there is no foreshortening at all, it means that the cylinder has a length of 0 - something that is clearly not true based on what we can see on the page. Ultimately this results in the different visual cues present on the page contradicting one another.

As a rule, you shouldn't be ending up with situations with zero foreshortening. The only situation where a vanishing point actually goes to infinity (allowing for something like this to happen) is when the cylinder is aligned in a very specific orientation that runs perpendicular to the viewer's line of sight. Since we're rotating our cylinders freely, we can assume this will not occur.

Moving onto your cylinders in boxes, here we've got lots of foreshortening and convergences, so that's great to see. As a whole you've done a good job, there's just one situation that I want you to be aware of, as it can sometimes happen when we're not paying enough attention. If you look at cylinder 193, you'll see that this one came out looking very squished, but the line extensions look to be pretty close to correct, with the blue contact lines being just a little bit off.

There is a piece missing here, however - you didn't check for the ellipses' minor axes. The minor axis passes through the narrower dimension of the ellipse, cutting it into two equal, symmetrical halves. Looking at these ellipses, the minor axis would be more similar to the blue lines passing through their centers, even though the minor axis should be passing through the "spine" of the cylinder, in a totally different direction. That's why this cylinder is incorrect, even though the line extensions you've drawn don't entirely suggest that it is.

Generally speaking you've done a good job across most of the others - you're developing your ability to construct boxes that feature a pair of opposite faces which are roughly square in 3D space. That's exactly what this exercise is about, and by testing the line extensions of our ellipses (and whether they converge towards the box's own vanishing points), we can test whether those ellipses represent circles in 3D space. If they do, then logically the planes that contain them must also represent squares. So, the more we work on drawing boxes that will ultimately end up with line extensions that all line up correctly, the better we get at drawing boxes of this specific nature.

So - you definitely have some things to keep in mind, but overall you're doing well. I'll go ahead and mark this challenge as complete.