Crematogaster missouriensis Emery, 1895

Shattuck, Steve & Cover, Stefan, 2016, Taxonomy of some little-understood North American ants (Hymenoptera: Formicidae), Zootaxa 4175 (1), pp. 10-22 : 15-16

Crematogaster missouriensis Emery, 1895 , new status

Crematogaster victima missuriensis Emery, 1895: 287 . 2 worker syntypes, Missouri (specific locality not given) (T. Pergande) (Museo Civico di Storia Naturale, Genoa, Museum of Comparative Zoology) [http:// mczbase.mcz.harvard.edu/guid/ MCZ: Ent:20816]. Raised to species.

Crematogaster minutissima thoracica Creighton, 1939: 138 (junior primary homonym of C. thoracica Santschi , replacement name C. smithi Creighton, 1950: 205 ). Holotype and paratype workers, Miller Canyon, Huachuca Mountains [31°25′N 110°16′W], Arizona, 6000ft., (W.M. Mann) (Museum of Comparative Zoology) [http://mczbase.mcz.harvard.edu/guid/ MCZ: Ent:23667]. New synonym.

Crematogaster minutissima smithi Creighton, 1950: 205 (new replacement name for C. thoracica Creighton, 1939: 138 ). New synonym.

Crematogaster minutissima and its associated taxa C. missouriensis and C. smithi are the northernmost representatives of a neotropical species group within Crematogaster . Longino (2003) reports the occurrence of C. minutissima in Costa Rica, but here we are concerned solely with the forms reported from North America north of the Mexican border. The relationships between these forms and the Caribbean and Central American representatives of this complex deserve future investigation.

Crematogaster minutissima was described by Mayr (1870) from specimens collected in “ Texas.” Crematogaster missouriensis was described by Emery (1895) as a subspecies of C. victima , apparently on the recommendation of Pergande (as the original description attributes the name to him). Creighton (1939) correctly recognized that C. missouriensis was actually more similar to C. minutissima than C. victima and regarded it as “northern race” of C. minutissima , maintaining it as a subspecies but transferring it from C. victima to C. minutissima . Crematogaster smithi was described by Creighton (1939) using the name C. thoracica , not realizing that that name had previously been used by Santschi (1921). Creighton (1950) corrected this oversight by proposing the replacement name C. smithi , the name it has been known by since.

Since the original describers were dealing with extremely limited material, the first meaningful assessment of these three taxa is Creighton (1939). Creighton regarded C. missouriensis and C. smithi as subspecies of C. minutissima , and separated all three using differences in mesosomal sculpturing and propodeal spine length. There matters stood until Mark Deyrup, drawing on extensive field experience in Florida, realized there were two species of yellow Crematogaster present in the state. One is found throughout Florida . It is a forest ant, nesting in rotten wood, nuts, pine bark at the base of large trees, and in thick litter. Colonies are diffuse, polydomous, highly polygynous, and the queens are small (4-4.8 mm long). The second occurs only in northern Florida, and creates inconspicuous nests in soil (sometimes surmounted by a small crater) in open or semi-open habitats. Colonies are monogynous, monodomous, and modest in size, often consisting of 100- 200 workers. Queens are distinctly larger, 5-6 mm long. Collecting by the second author, Mark Deyrup , and others has revealed that precisely the same situation prevails throughout most of the southeastern and south-central states. The small queen species occurs from North Carolina south, and then west along the Gulf Coast to Arkansas, Missouri, Oklahoma, and central Texas . It maintains its distinctive nesting habits and population structure throughout. Likewise, the large queen species occurs from North Carolina and Tennessee south to northern Florida, and west to Iowa, Oklahoma, and Texas. It too is consistent in its ecology and colony structure.

The small queen species is clearly Mayr’s C. minutissima . Mayr gives the length of the syntype queen as 4.2 mm, which is decisive. A specimen of the original Pergande collection of C. missouriensis from Missouri is present in the MCZC (see Creighton, 1939). Its longer propodeal spines and lack of longitudinal rugae on the promesonotal dorsum matches Creighton's concept of C. missouriensis , separates it from C. minutissima , and therefore establishes it as a worker of the large queen species. It is here considered a full species in its own right and not a subspecies of C. minutissima .

This leaves us with Crematogaster smithi . The worker of this ant was diagnosed as having a relatively smooth and shiny mesosomal dorsum compared to the dull surface supposedly found in close relatives, and as lacking longitudinal rugae on the pronotum (or when present they are lateral and wavy). In related species these rugae are well developed and more central on the pronotum. However, Creighton noted that the rugae in C. missouriensis are variable and can also occur laterally, but in these cases the rugae are more prominent than those found in C. smithi . Also, in dorsal view the petiole of C. smithi was described as diverging posteriorly and with the widest point near the insertion of the postpetiole while in C. minutissima and C. missouriensis the petiole is quadrate with subparallel sides, the widest point (when noticeable) being near the center. Finally, Creighton noted that the propodeal spines were short and in other respects this taxon was very similar to C. minutissima .

An examination of recent collections has convinced us that the morphological characters cited by Creighton are all more variable than he suspected and that there is considerable overlap with C. missouriensis and C. minutissima . In addition, recent collecting has narrowed the distributional gap between these forms. C. smithi was described from specimens collected in Huachuca Mtns. in south-central Arizona . Since then, it has been found widely in Arizona, and two additional important collections have been made: one in the Davis Mtns of west Texas (including queens) and another from Roosevelt Co. , New Mexico (both in MCZC). The Davis Mtns. collection clearly matches Creighton’s concept of C. missouriensis and the Roosevelt Co. specimens are a good match for his concept of C. smithi . This may seem confusing, but natural history comes to our rescue here. Queen size, colony structure, and ecology of C. smithi all closely match those of C. missouriensis .

Our assessment of the evidence is that there are only two species: C. minutissima , and C. missouriensis . In the eastern and central parts of their range, where C. minutissima is present, the workers of C. missouriensis are fairly consistent morphologically and separable from those of C. minutissima . In the absence of C. minutissima (not present west of the Hill Country in central Texas), western populations display greater variation in worker morphology, a situation reminiscent of classical character displacement. However, the western populations are firmly linked to C. missouriensis by queen size, colony structure, and ecology. C. smithi , therefore, is best regarded as a synonym of C. missouriensis .