Quick & Dirty Division

Dividing things visually is pretty critical for drawing- a teacher of mine used to go so far as to say "Drawing is division".  When doodling in your sketchbook, it's easy to start with a head, maybe add a torso, and if you feel like it (and there's still room on the page) you can throw in some legs.  But when you start thinking compositionally, you need to be able to say "The figure goes here".  That is to say, it needs to fit within a certain spot on the page or the greater composition in order to create a good whole.

When it comes to that point, it's critical to start thinking in terms of division.  Unfortunately, humans are pretty terrible at dividing anything other than in half- and they're not very good at that over large distances.  Head lengths, a common proportional device, fail us pretty quickly.  Unless the figure is 8 head-lengths, which is convenient for division, we can't divide our composition space into some weird number of head lengths that most people are (7.5, 6, 6.5).  Proportional dividers and grid transparencies can help with this, but sometimes they're too cumbersome to deal with- especially in quicker poses or when working on larger surfaces. 

Plus, all of this is assuming a nice standing figure with little foreshortening.  If the model is crouching or reclining, head lengths pretty quickly become useless- even with the proportional divider.  In such a case, or if the pose is quick or you don't want to use too many aides in your drawing, you can use this technique to quickly find a halfway point of something.  It works on both the model and on your drawing.  All you need is a straightedge, which a relatively straight pencil, brush, piece of charcoal, or knitting needle can function as just fine.

1. Choose a top and bottom point.  If it's your drawing and it's in the beginning stages where there aren't many marks on your page to grab on to, make a small mark at the top and bottom.
2. Estimate a halfway point.  Don't labor about it- just take your best guess.  It's most likely going to be wrong, but it will help us quite a bit nonetheless.  If you're measuring on your drawing, put a very faint mark where you think halfway will be.  If you're looking at the model or some other subject matter, just choose some landmark on the subject.
3. Extending your arm all the way and locking your elbow, take your straightedge and put the end point at your estimated halfway point.  Slide your thumb down so it lines up with the bottom mark that you made or decided upon.
4. Keeping your arm extended and your elbow locked to minimize forward and back movement, shift your arm up so the end point is now at the top mark.
5. Unless you made a very good estimate, your thumb should be above or below your estimated halfway point.  Here's the key- halfway between your thumb and the estimated halfway point is the actual halfway point.  Since it's a much smaller distance to work with now, you can now estimate a new halfway point based on this new information, which will be significantly more accurate.  If you're looking at the subject, find a new landmark to hold on to mentally.  If you're measuring on your drawing and there's nothing on the page to grab on to, make a new mark on the piece
6. Keeping the end point of your straightedge aligned with your top mark (or putting it back there if you used your straightedge to make a mark), slide your thumb the new & improved estimated halfway point.
7. Because we're human, it's good to check out our estimations.  Small shifts in our position can make a big difference, as can things significantly above or below our view point.  Further, ssuming something is correct that is slightly off can be the death of proportions.  The beauty here is that to check things out, you simply have to repeat the process. Go back to step three and repeat until you're satisfied. On a quick drawing I might do this process only once, but on longer drawings I might repeat it several times.

It's important to note that this process- or any measuring technique- will only get you so far.  They sound really good on paper, but they are never as accurate as they might seem.  Most people go through a phase when they begin to understand such techniques where they become overly dependent on them.  They try harder and harder, but still things don't line up.  While there is a certain degree of meticulousness that will bring results, it doesn't seem to be a silver bullet.  It's best to supplement this kind of measuring with other kinds of proportional systems, especially constructive systems.  I tend to use such measuring just to get me in the ballpark, or to cross-check other proportional estimates.