Yeesh is right! You certainly worked hard on completing this challenge, and overall you did a pretty good job. While there are certainly areas to work on, the growth over the set was fairly considerable.

Starting with your cylinders around arbitrary minor axes, I'm very pleased to see that you were conscientious with the execution of every single mark, apoplying the ghosting method with great care, and in your analysis of the results. While mistakes are inevitable, it is critical that we identify them, so we can actually learn from them and use them as stepping stones towards improvement. I'm also very happy to see that you experimented with a variety of different orientations, sizes, and rates of foreshortening.

This last one actually helped to identify one issue that I specifically look for in this exercise. It comes down to the fact that when drawing our cylinders, the ellipses on either end demonstrate two different kinds of shifts in their properties. On one hand, there's the standard "scale" shift where the farther end is to be smaller than the closer end. The other is the "degree" shift, where the far end is to be wider than the closer end, as explained here. I actually didn't see as much of the latter earlier on in the set, but as you moved through it, it started to show up more frequently.

Both of these shifts are aspects of foreshortening - when the far end is much farther away from the closer end, we see a more significant shift in scale and in degree. These both serve as visual cues, telling the viewer that the cylinder is longer (specifically when it's moving more through the depth of the scene, rather than across our field of view). There are however cases when a student will forget this relationship, and where the shift in scale may be minimal, but the shift in degree is more significant (or vice versa). This makes the cylinder feel "off", though the viewer won't be able to pinpoint why.

We see this issue come up in a few cases - like cylinder 98, where there's almost no scale shift, but a more notable degree shift. In most cases though, you did a pretty good job of keeping them consistent. That said, do remember that there's always going to be at least a little bit of both kind of shift.

Moving onto your cylinders in boxes, you definitely struggled a lot more with this in a number of ways. For example, you weren't always able to draw ellipses that touched all four edges of the planes they were being placed within, along with some cases where you struggled with the construction of the boxes themselves. Perhaps you may not have been keeping up with those exercises as parts of your warmups - so make sure you do moving forward.

You did however improve on both these fronts (although on the 99th you still didn't quite get the ellipse to fit entirely within the plane). You also improved in terms of the general purpose of this exercise - which, though I purposely don't make it clear, is all about learning to construct your boxes such that the proportion of two opposite faces is roughly square. We achieve this similarly to how we use the box challenge to improve your ability to intuitively construct boxes with more consistent convergence - we use the line extensions to analyze our mistakes, and then adjust our approach in the next attempt, repeating this over and over and gradually shifting how our brain handles spatial reasoning.

In this case, we add to those line extensions those brought to the table by the ellipses - the minor axis and the contact points. These additional lines will only actually align to the box's vanishing points when the ellipse itself represents a circle in 3D space that is laid flat against the plane it is drawn within. If that ellipse is a circle, then the plane containing it must be a square. So, by going through the extensions over and over, we gradually sharpen our instincts in terms of eyeballing the correct proportions.

And in this regard, you've definitely shown significant improvement. All in all I am quite happy with your results. As such, I'll go ahead and mark this challenge as complete.