A feel for big numbers, with grains of salt 

salt grains 

Counting...
(with a grain of salt, say ±50%)
a grain of salt  One  10^{0}  1. 
some grains  ten  10^{1}  10. 
a tiny pile  hundred  10^{2}  100. 
a pinch  thousand  10^{3}  1,000. 
1/4 teaspoon  ten thousand  10^{4}  10,000. 
tablespoon  hundred thousand  10^{5}  100,000. 
1/2 cup  million  10^{6}  1,000,000. 
ten million  10^{7}  10,000,000.  
hundred million  10^{8}  100,000,000.  
a bathtub  billion  10^{9}  1,000,000,000. 
ten billion  10^{10}  10,000,000,000.  
hundred billion  10^{11}  100,000,000,000.  
a classroom  trillion  10^{12}  1,000,000,000,000. 
Perhaps this would be better done with sugar, as it tastes better...
With a 1mm grain size, it would be something like:
sugar grains 

With 1/4 billion people (1/4 bathtub of grains) in
the US... How much are under 14 years old? Unemployed? How fast
is salt being added by birth? Removed by death?
(Lots of
stats available...)
With a room filled to the ceiling with a trillion grains, when you
open the door and paw at it with cupped hands, how much/fast do you
move? With a shovel? A bucket?
(The US
budget is a trillion and a half dollars. How much per day?
Second? How much is Medicare? Social security? Education? Space?)
Salt is fascinating stuff. Roman legions were payed in it, cities taxed, evaporating saltwater for it was a crime. Politics, rebellions, required for life, preservation of food. Perhaps someone has done a web page...
Comments encouraged.  Mitchell N Charity <mcharity@lcs.mit.edu> 
History: 2003Feb03 Fixed 2 links. 2000Oct17 Repointed "stats" link to something similar that works. Thanks to a reader. 1997.May Created.
Some details about this page:
The back of my envelope... I'm using "Morton Salt" (table salt). First, some quick and dirty work... The grains measure about 1/2 mm, which gives 20/cm 8000/cm^{3} 8 x 10^{9}/m^{3} 10^{10}/m^{3} Its box is a cylinder 6 x 10^{4} m^{3} (8cm d, 12cm h), and masses 0.737 kg, so thats a density of about 1.2 kg/m^{3} (1.2 g/cm^{3}). So 10^{10} kg/grain, and 5 x 10^{6} grains in the box. More carefully measuring the grains, I get 0.5 + 0.05 (about 1 grain uncertainty in 5mm). Upping the sample to 10mm didnt seem to help narrow the uncertainty (due to grain size variations). With this big an uncertainty, I'm not going to worry about how tighly packed the grains are. thus 20/cm (+2) 6000 to 11000 /cm^{3} 0.6 to 1.1 × 10^{10} so 10^{10} (+10%40% for measurement uncertainty) which gives worstcase percent errors of +70% and 10%. So, probably an overestimate, possibly by a factor of 2, but not much of an underestimate. So... grains/m^{3} 10^{10} × 0.6 to 1.1 kg/grain 10^{10} × 1 to 2 grains/kg 10^{10} × 0.5 to 1 Ok, what are some accessible volumes. teaspoon 4.9x10^{6} m^{3} tablespoon 1.5x10^{5} m^{3} cup 2.4x10^{4} m^{3} gallon_{US liquid} 3.8x10^{3} m^{3} foot^{3} 0.028 m^{3} bathtub ~0.2 m^{3} (~8 ft^{3}) classroom ~200 m^{3} (10x8x2.5 m) So teaspoon 3 to 6 × 10^{4}, or tablespoon 0.9 to 2 × 10^{5} cup 1 to 3 × 10^{6} bathtub 1 to 2 × 10^{9} classroom 1 to 2 × 10^{12} Ok. Time for a reality check. A big pinch looks about 10^{3} grains (by dividing it in half a couple of times and then counting). So a 1/4 teaspoon seems 10^{4} or two. Which sortof fits. (_This_ time it fits. The first time it didn't. Turned out I had mangled the object vs # grains paragraph. And in tracing it down, I would have done a lot better to let the numbers lead me, rather than having an emotional stake in their fitting. A lesson.) I could do this more carefully, but for now... Call it a 1/4 teaspoon for 10^{4} a tablespoon for 10^{5} a 1/2 cup for 10^{8} a bathub for 10^{9} a classroom for 10^{12}.
I should address the nonlinear error "feel". 50% seems a bigger deal for big numbers than for small ones.