Starting with your cylinders around arbitrary minor axes, you've certainly been rather fastidious about checking your alignments, and for the most part it has certainly helped you improve upon them as you worked through the set. There are however a few key issues that I want to point out:

  • The assignment section of this challenge specifically requested that you include lots of variation in the rates of foreshortening you were using. You appear not to have done that.

  • Taking the previous point a little further, there are a number of places where it seems like instead of being mindful of having your side edges converge towards a shared vanishing point, there are definitely some where you've kept them roughly parallel on the page, with no sign of any intent to have them converge. This is a common issue for students, but it's one to be aware of - when you're just rotating forms randomly in space, you should always be mindful of having their sets of parallel lines converge in 2D space. Reason being, those lines will only be parallel on the page (with a vanishing point at infinity) when the lines in 3D space run perpendicular to the viewer's angle of sight. If they're at any other orientation, then having no convergence there will tell the viewer that there is no distance between the ends of the cylinder, resulting in a visual contradiction (since we can clearly see that there is distance between those ends).

  • It does seem like your ellipses still do have room for improvement (mainly in tightening them up as you draw through them two full times before lifting your pen - something you don't actually do entirely consistently as you should be). This is honestly pretty normal, because ellipses are frankly, difficult. But be sure to keep using the ghosting method to prepare your arm, and to draw using the whole arm from the shoulder. It's easy to slip into drawing these from the elbow.

Continuing onto your cylinders in boxes, here I'm pretty pleased with your work. Some of your boxes come out a little wonky, like 94, but by and large you're doing a good job of minding those convergences, and of extending the analysis to include the ellipses' own lines (like the minor axis and contact point lines). In doing so, you've done a good job of further training your ability to estimate the construction of boxes that feature a pair of opposite ends which are proportionally square.

This gets developed largely because we're working towards getting those line extensions to converge towards shared vanishing points - when the ellipses' lines converge towards the box's vanishing points, it means those ellipses represent circles in 3D space. That of course would also mean that the plane enclosing them would represent a square in 3D space. As we continue to try to adjust the boxes to get better results on the extensions, we train this skill often without being entirely aware of it.

Fortunately, it's an important skill, and one that'll serve you well in the next lesson. So, I'll go ahead and mark this challenge as complete.