I also categorize the material by subject. This is treacherous but I expect ultimately useful. Munsell's printing was decidedly local and often mundane. There are no Gutenberg Bibles here although interspersed with obsequies are such items as "Stories of the Early Settlers in the Wilderness,..." [Priest], "Report on the Phrenological Classification of J. Stanley Grimes" and "A Treatise on Roads, their History, Character and Utility..." The categories I employ are: local, science, almanac, history, education and periodical. They are not chiseled in stone.

A spreadsheet of these two decades is attached. 1830's and 1840's

The first impression that emerges is that the number of pages accurately predicts survival. There is also a rising survival rate by decade. In the 1830's, for pamphlets up to 16 pages, a single copy survives for every 500 printed and in the 1840's 1 for every 416. Pamphlets printed by Munsell in the 1830's, with from 17 to 71 pages, survive at the rate of 1 for every 232 printed. In the 1840's the survival rate improves to 1 in 172. When the page count reaches or exceeds 72 the survival rate for material printed in the 1830's is 1 in 149, in the 1840's 1 in 78. All other issues aside size is a clear predictor of survival.

Survival Rates By Page Count
	1830's	1840's
1-16 pgs	1 in 504	1 in 419
17-71 pgs	1 in 230	1 in 172
72+pgs	1 in 150	1 in 78

Classifying material by category should be useful but isn't so far.

1830's Material By Category				1840's Material By Category
	1-16pgs	17-71pgs	72+pgs		1-16pgs	17-71pgs	72+pgs
local	1 in 492	1 in 144	1 in 100	local	1 in 585	1 in 120	1 in 118
Science	1 in 1060	1 in 153		Science	1 in 40	1 in 264	1 in 87
Almanac		1 in 500		Almanac		1 in 2636
History		1 in 652	1 in 103	History	1 in 23	1 in 151	1 in 20
Periodical			1 in 750	Periodical			1 in76
Education			0 in 2000	Education	1 in 400		1 in 94
Literature				Literature	1 in 1000	1 in 211	1 in 646

There are two obvious problems. The total number of items for these two decades is less than 300. Dividing into two decades and so many categories isn't practical yet. When I add the 1850's and 1860's the number of applicable items will increase to almost a thousand and the statistics should make more sense. The other issue of course is that I have certainly mischaracterized some of the material. Hence the categories will certainly change and the characterizations as well and with these changes and an increasing number of records more will be known.

For the moment this is simply a work in progress.