Despite your struggles, I honestly think you've done a pretty great job. One thing I should make clear is that ellipses are hard. It's not just you. Drawing functional, good-enough-ellipses is something we can achieve with some practice, but the last mile from that to really good ellipses simply takes far more mileage than we're able to cover in the scope of this course. That's precisely why the last few lessons after this allow for (and encourage) the use of ellipse guides.

All that said, your ellipses are a little loose, their alignments are a little off here and there, but in terms of what I'm expecting from students at this point, they're solidly done. I'm also pleased to see that you've been very fastidious in checking your alignments afterwards, and that you incorporated cylinders with a variety of lengths and rates of foreshortening.

One thing I do want to warn you against are the ones where you kept the scale of the ellipses from one end to the other roughly the same. For example, number 31 and 33 are good examples of this. In essence, we can look at these and say that looking at the shift in scale, there is no foreshortening being applied to this cylinder.

Of course, if we look at the shift in degree (where the far end is wider than the closer end), we can in fact see some foreshortening - which already suggests that there's a contradiction here, with one of the manifesting signs of foreshortening telling us that there isn't any, and the other telling us that there is. That's one issue.

The other issue is that there's a very specific circumstance in which we would see no scale shift from one end to the other. This would only occur if the side edges' vanishing point is "at infinity" (as discussed way back in Lesson 1). One misconception is that one can just choose to draw with vanishing points at infinity - but that's not the case. It's based on how that cylinder is actually oriented in space. If that cylinder is running perpendicular to the viewer's angle of sight, then the vanishing points of the side edges will be at infinity, and we'll see no scale shift from one end to the other.

If the cylinder is oriented in any other way, then there has to be a scale shift, otherwise you'd be telling the viewer that there's basically no physical distance between the ends, and therefore the cylinder has a length of zero. Of course, with both 31 and 33, we can physically see that they have a non-zero length.

Long and complicated story short - be very careful when you decide to eliminate the convergence on a set of lines that are parallel in 3D space. Either those lines are running perpendicular to the viewer's angle of sight, in which case you must keep them parallel on the page, with a vanishing point at infinity, or they're not running perpendicular to the angle of sight and you cannot keep them parallel. You might give them a veeeery gradual, minimal convergence, but it has to be there.

Continuing onto your cylinders in boxes, overall you've done a good job here both in constructing your boxes/cylinders and in checking those line extensions, and as a result I can see definite improvement in your ability to estimate the proportions of a box in order for it to have a pair of opposite faces which are square in 3D space. You're certainly not perfect in that regard, but that's not what's expected. Instead, I can clearly see that those line extensions are getting closer and closer to converging - and therefore, the ellipses' lines are lining up more closely with the box's own vanishing points. This of course tells us that the ellipses represent circles in 3D space, and therefore the planes enclosing them represent squares.

Now, that aside, I am a little, very so slightly concerned about your linework. The edges of your boxes themselves don't necessarily feel like they're executed as confidently as they could be, with the slightest little wobbles and hesitations here and there. I'm also not seeing the usual signs of the ghosting method (the little points that would define the start and end of your lines), which leads me to question whether you're using it as consistently as you should.

Remember - the ghosting method is at the core of this course. It forces us to think and plan and weigh the relevance of each and every mark before execution, and ultimately encourages a confident, smooth, consistent stroke at the end. Make sure you're applying it to each and every mark, without exception.

Anyway, I'll go ahead and mark this challenge as complete. You're doing a good job, so keep it up.