Starting with the cylinders around arbitrary minor axes, these are largely looking quite well done. You've done a great job of analyzing and identifying the correct minor axes in your corrections, your ellipses are looking evenly shaped and confidently drawn, and your cylinders capture the appropriate relationship between the shift in scale from one end to the other, and the shift in degree. That is to say, I'm not seeing many inconsistencies where the shift in scale is more dramatic and the shift in degree is more limited. You keep them consistent, maintaining a cohesive sense of foreshortening in the vast majority of cases.

Moving onto the cylinders in boxes, it took me a second to grasp what you meant in your comment, but looking more closely it became clear. The answer, fortunately is actually quite simple. The characteristics we use for our ellipses' placement/alignment - that is, the alignment of the minor axis and the lines defined by the contact points will only line up correctly to the box's vanishing points if the ellipse actually represents a circle in 3D space that rests upon the surface of that box. Now, if the ellipse represents a circle, then the plane enclosing it would ostensibly be proportionally square.

Now here's the actual answer to your question - if that plane is in fact not a square in 3D space (which will depend on its proportions, the position of the vanishing points, etc.) then the ellipse's own lines will not align correctly, and the ellipse won't meet the criteria to represent a circle in 3D space. To put it more simply, if those lines don't align to the vanishing points, then the ellipse isn't a circle, and the plane isn't a square.

This exercise is, at its core, more about the box than the cylinder. It's all about training students' ability to roughly draw boxes that have two faces which are proportionally square. Just like the box's own line extensions, we add the ellipses so we can test those proportions. As we identify the mistakes and make adjustments to bring them more in line with the vanishing points, we gradually train our instincts, making it easier to eyeball these proportionally square boxes correctly. This does however require us to actually draw ellipses that fit snugly within the face of the box, rather than allowing the ellipses to fall short of an edge.

Despite the fact that you didn't always get the ellipse to touch the edge correctly, you did still improve a fair bit in your estimation of those proportions, and I can see you nailing them with much greater confidence later on in the set. All in all, I'm still quite pleased with the results.

So! With that settled, I'll go ahead and mark this challenge as complete.