forms similar to the first-known species 

 persist unchanged today. The first 40 mil- 

 lion years of their fossil record shows little 

 change, but during the past 100 milhon 

 years, hundreds of genera diversified. In 

 each well-studied case, however, the new 

 species appeared relatively abruptly, then 

 remained remarkably stable over vast 

 amounts of time. 



Our research strategy was to begin with 

 hundreds of samples of modem bryozoan 

 skeletons, and assign each to various spe- 

 cies by measuring their structures and 

 ranges of variation — a tedious statistical 

 procedure. Next, we tested the reality of 

 the "species" our measurements had de- 

 fined. Would they match up with species 

 defined by breeding experiments and tests 

 of their genetic biochemistry? We foresaw 

 three possible results. First, as some skep- 

 tics suggested, skeletal features might be 

 too uninformative to be useful in distin- 

 guishing species. Second, environmental 

 differences might produce very different 

 external characteristics in colonies of the 

 same species, creating "false" species. 

 The last possibiUty — the one we suspected 

 was true — was that skeletons are rehable, 



environmentally stable definers of bry- 

 ozoan species, equivalent to genes in their 

 precision. But to support our case, we 

 needed a very high level of agreement in 

 the species markers, by several different 

 means of measurement. 



We first studied three distantly related 

 genera of bryozoans that live in the San 

 Bias Islands and elsewhere along the Car- 

 ibbean coast of Panama, each very differ- 

 ent in its skeletal complexity. We would 

 compare them with fossil bryozoans based 

 only on skeletal measurements. Among 

 the features we measured were the dimen- 

 sions of the apertures, the shape of the cal- 

 careous modules that encase the body, and 

 the comma-shaped stiaictures (the avicu- 

 laria) that are used to protect the oral open- 

 ing. Some species' skeletons had rela- 

 tively simple shapes that were easy to 

 describe in ten measurements, while the 

 more complex types had additional struc- 

 tures that required as many as forty. We 

 needed to know whether the number of 

 traits we measured was influencing our re- 

 sults. In all cases, we found the statistical 

 differences in the measurements of 

 twenty-two species held up, whether 















based on a few anatomical features or 

 many. 



Next, we needed to check on the stabil- 

 ity of the species in differing environ- 

 ments. We collected colonies from differ- 

 ent reefs and raised their offspring in the 

 shallow seawater adjacent to the Smith- 

 sonian's laboratory in the San Bias Is- 

 lands. After we had successfully raised 

 two generations there, we began to study 

 the offspring. Nearly 500 had grown big 

 enough for morphological analysis before 

 the experiment was terminated by a hurri- 

 cane. The results were clear. All offspring 

 in all three genera closely resembled their 

 parents, despite having been transplanted 

 to a new environment. No false species or 

 environmental variants appeared. 



We then used a biochemical procedure 

 called protein electrophoresis to study 

 variation in enzymes that are coded by the 

 bryozoan's DNA. This is a relatively old 

 and not particularly sensitive technique 

 that first came into general use during the 

 1960s, and has since been supplanted by 

 DNA comparisons. But electrophoresis 

 has the advantage of rapidly and cheaply 

 screening genetic variation in large num- 

 bers of animals. We examined more than 

 400 colonies of eight species. Again the 

 results were unambiguous: no genetic evi- 

 dence for undetected, or "hidden," species 

 and clear genetic differences between all 

 species tested. So far, the numbers pointed 

 overwhelmingly in a single direction: we 

 had the abihty to detect true bryozoan spe- 

 cies in the fossil record from their calcare- 

 ous skeletons alone. 



Finally, we extended the study to in- 

 clude more than one hundred Caribbean 

 and western Atiantic populations of the 

 genus Stylopoma, which is one of the three 

 bryozoans that we had first looked at. We 

 chose this genus because it is abundant 

 today, it has many different species, and its 

 fossil record seems complete enough to 

 help reconstruct the evolutionary relation- 

 ships of all known species in the genus. 



Fossil skeletons of the bryozoan 

 Metrarabdotos auriculatum,/oM/z£? in 

 the Dominican Republic, show how 

 little the species changed over eons. 

 The Pliocene species, left, is 3.4 

 million years old and the Miocene 

 species, far left, dates fivm 7.3 

 million years ago, yet their structural 

 details are virtually identical. 



Photomicrographs by National Museum of Natural History: 

 SEM Laboratory 



< 



o 



o 



3 



57 



