Applying modern analogs to understand the pollen content of coprolites

Most scientists working with coprolites from archaeological contexts assume that human fecal specimens reflect the mixing of the pollen ingested during the period in which the contribution to the coprolite, 19 to 37.5 h, was ingested, that the amount of pollen in a fecal sample directly reflects the amount of pollen originally ingested during that interval, and that differences between the amounts of pollen in different fecal specimens reflect differences in the quantities of pollen ingested at different times. These assumptions were tested and found wanting in an experiment in which two persons sequentially ate separate quantities of 15 pollen types in meals over a four-day interval. The pollen was retrieved and analyzed from feces produced during those four days and five days of subsequent fecal production. Pollen ingested first appeared in relatively small amounts, usually the day after it was ingested. Its concentration per gram of sample then increased rapidly and remained high over a one to three day interval relative to the amounts in previous and subsequent fecal specimens deposited. When pollen concentrations declined some pollen was retained in the gastrointestinal system and much lower concentrations per gram of sample of each type continued to appear in fecal samples for several days. These relatively low pollen concentrations appeared in fecal samples approximately twice as often as did higher concentrations. Our results indicate that comparatively high pollen concentrations can be used to determine that a given pollen type was ingested, but comparisons between pollen concentrations of the same pollen type in different fecal specimens or between different pollen types in the same fecal specimen, cannot be used to determine whether different amounts of pollen were ingested, or what was the relative amount of each ingested. Because pollen concentrations per gram of sample varied widely with time since ingestion, percentages of given pollen types did not occur in predictable patterns and could actually increase as the concentration of the pollen type decreases. Hence, percentages should not be used in coprolite pollen analysis. The experimental results also suggest that variations in the pollen content of different portions of a coprolite are meaningful only in terms of the overall pattern of a sequential group of coprolites.