Starting with your cylinders around arbitrary minor axes, I definitely get the impression that earlier on in this section, you weren't necessarily putting as much time into each mark as you could have, resulting in somewhat sloppy linework. I can see signs that you were applying the ghosting method for the minor axis line, but not necessarily for the side edges (which at times did end up being wobblier).

That said, after the first 20 or so, your cylinders definitely ended up becoming more confidently drawn, although this does vary throughout the set. Overall I do kind of get the impression that you didn't necessarily invest as much time as you could have into these cylinders, and while there is improvement, I suspect you can do better.

More importantly though, there was a pretty significant part of the instructions - specifically the bolded part of the assignment section - which you appear to have missed. You did not include any variation in foreshortening to your cylinders, and from the looks of it, you appear to have avoided any foreshortening at all, at least in regards to the shift in scale we get where the closer end is always smaller than the farther end (based on the basic rules of perspective).

Foreshortening manifests in two fashions - there's the scale shift I just mentioned, and the shift in degree where the farther end is wider than the end closer to the viewer. Both of these work in concert - meaning that if the cylinder is longer and there is more distance between the ends, the shift in scale will result in a much smaller far end, and the shift in degree will result in a much wider far end. Conversely, when the cylinder is shorter and that distance is more limited, there will be less shift in both regards.

In your cylinders, you are including a shift in degree which suggests that there is some length to the cylinders, but if we look at the scale shift, the viewer is being told that there is no distance from end to end, that the cylinders have a length of 0. Of course this is contradicted by the actual distance we can see on the page itself, so the viewer is left feeling that something is off.

This exercise should definitely feature cylinders with both dramatic foreshortening and those with shallower foreshortening, but to have them with no foreshortening is entirely incorrect. The only situation where the foreshortening is completely nullified is when the orientation of the cylinder is perpendicular to the direction the viewer is looking - which is what, as described in Lesson 1, results in a vanishing point "going to infinity". Since these cylinders are all freely rotated in totally random orientations, we can pretty much assert that this orientation wouldn't really occur.

Moving onto your cylinders in boxes, your work here definitely shows a lot more overall progress. You started pretty weak with some seriously wonky boxes, but you pick up pretty quickly and for most of them you're showing far improved convergences, and your estimation of the correct proportions for the boxes to keep the faces where your ellipses will be drawn square in 3D space definitely improve a great deal.

I did notice one thing - you don't seem to be checking your minor axes as carefully as you ought to - I caught these where the minor axes were way off, resulting in really squashed cylinders, but you don't appear to have actually identified the issue yourself.

Anyway, I'm pretty happy with your cylinders in boxes, but your cylinders around arbitrary minor axes leave a lot to be desired. I'm going to assign some revisions below.