That's a great question, William.  I had an answer in mind, but after thinking about it a little, I'll say this instead:  I think that mass analysis is a pretty rudimentary gauge for assessing production technology or reduction staging because other factors can affect debitage size distributions besides knapping behavior  These include post-dep geologic forces, cultural formation processes, and variance in the acuity of individual field technicians to systematically collect flakes from a screen.  I'm not saying anything here you don't know already.  The interesting thing is that all of these non-cultural factors have their greatest impact on small flakes rather than larger flakes.  For example, wind will sort small flakes from a deposit and field sorters will overlook more small flakes than large flakes.  So in a way, looking at 1/4-in. flakes and larger gives you more signal and less noise. 

There is a catch though.  The size of the parent piece is also important.  Looking exclusively at larger flakes may be workable at lithic source areas where large cores and bifaces are being reduced, but i don't think 1/4-inch and larger flakes are going to sufficiently capture the sort of reduction usually seen at camps where tools are finished, reworked, and resharpened.  For this reason, I think it's critical to include the 1/8-inch fraction.  I'm interested to hear what others say about this topic. 

William,

First, I understand that >1/8" mesh is common, but I cannot support it. But I agree with Doug (above) that mass analysis is rudimentary. And all stages of core reduction results in microflakes, and we will never get them all. Mass Analysis will never solve this problem. But I like mass analysis, with a caveat.

Mass Analysis will not, in my experience, answer any questions, but will raise many, and there is its value. This value should not be underestimated, though. It is a quick way to establish that there is "something funny here," but you must decide what that is. Science doesn't advance by "Eureka!" but by "that's funny..." Answering those "funny" questions is is our challenge, but at least the questions have been raised, and they may not have been even raised otherwise.

Some of those questions, as Doug notes, may be problematic for the very small flakes; but we take what we can get.

Please keep me informed as to your progress, it is a difficult problem, and I am interested in the way you solve this.

Mike Walsh

I am assuming you are looking at material already excavated. If that is the case, the even bigger problem is that debitage with different size distributions reflect different sorts of activities. For an easy example, core reduction leaves a different pattern than retouch/sharpening during use. The coarser the "grain" of your data, the coarser the picture of activity.

If your are planning your field work, then use 1/8 screens. I also took 1 liter "float" samples from every 10 cm excavation level for random grid units (other procedures for features, of course).  I used the macro, and  microscopic sort ( down to 1.5 millimeter) of the heavy residue to do a rough calibration of size fractions in the larger dry sample.

Nearly every site I worked on had massive rodent activity. A pocket gopher moves over a ton of soil per year, and a lot of that is laterally up to 2 meters. The worse problem was that the little buggers size sort by selectively moving small material to the surface, and digging under larger objects. If you get a Poisson distributed size fraction the result will be nearly all rodent sorting.

Just to chime in, and in agreement of the other replies- Ahler's mass analysis works for Ahler at the Knife River Flint quarries, but its use anywhere else is problematic because there are too many variables in lithic reduction, as Douglas noted.  If you don't have the 1/8th inch fraction, it's even less useful (I'd go so far as to say it's useless). If the collection is already excavated, do a technological analysis of the debitage that you do have; if you haven't excavated yet, use 1/8th inch screens-- and then do a technological analysis of the debitage, not mass analysis.  The appeal of mass analysis is that it is quick and easy, but like many quick and easy methods it won't provide useful data, particularly in your case.

I would like to thank everyone for their useful answers to date. I would like to clarify my question if I may. There are three general trends made evident by Ahler that appear to be applicable to any assemblage using this method. Namely, there is a decrease in the average weight of flakes, an increase in flakes retained in the 1/4" size grade, and a decrease in cortical flakes as reduction continues. Ahler's method is certainly not without problems (as are all lithic analytical methods), such as the interpretive value of cortex and the issue of mixed-assemblages. Some have tackled the problem of mixed assemblages with some apparent success using regression analysis on simulated mixed assemblages (e.g. Bradbury and Carr 2004), and while cortex is primary related to initial core size rather than "stage" it may inform on questions of raw material procurement. It is certainly beneficial to have an experimental data set derived from the same particular chert type in question or relatively similar. My question is thus geared towards how other researchers have used the general trends of mass analysis in the absence of the 1/8" size grade.

William, while Patterson's log analysis is simplistic and has its faults, I have used it on numerous Phase III excavations to get favorable results when comparing lithic assemblages with experimental lithic reduction results. I also wrote a paper comparing the size trends in the reduction of split cobbles, whole cobbles, and quarry materials. Unfortunately, several of the graphs and tables were lost during data corruption and I'm still trying to restore them.