The Cryptophagidae of Canada and the northern United States of America

CJAI 40 -- December 20, 2019
doi:10.3752/cjai.2019.40

Georges Pelletier & Christian Hébert

An Editorial Corrigendum has been published for this paper doi:10.3752/cjai.2019.40ed.

DNA Barcoding and neighbour-joining trees

The DNA barcoding methodology used in this publication was designed by Dr. Paul Hebert from the University of Guelph (Ratnasingham and Hebert 2013) and is explained in detail in Milton et al. (2013).
We submitted 76 specimens of Cryptophagidae to the Canadian Centre for DNA Barcoding at the University of Guelph. Most of these belonged to the genera Cryptophagus and Atomaria, including 12 paratypes of 8 recently described species, to confirm their identity as distinct species. The Search Data tool was used to search all DNA-barcoded specimens of Cryptophagidae in public records and our project’s databases. In order to build the genetic sequence barcoding, we selected specimens containing sequences from the mitochondrial gene COI-5P with the following criteria:

• Collected in Canada and the US between 1998 and 2018.
• For Holarctic or adventive species, European specimens available to compare DNA barcoding with Nearctic specimens.
• Specimens with images that can be identified at the species level included in the BOLD database.
• Minimum sequence length of 500 bp, giving priority to Full Length Barcode specimens of 640 bp or more.
• Barcode-compliant specimens.

We produce dendrograms with the Taxon ID tree engine using the nucleotide as a coding marker, the Kimura 2 Parameter model and the BOLD aligner that seems to be more relevant to discriminate species within large genera and to discriminate genera within Cryptophagidae. A dendrogram of clustered species can be seen in the Taxon ID Tree 1 (Supplementary Materials) for representatives of most genera of Cryptophagidae, Taxon ID Tree 2 (Supplementary Materials) for Cryptophagus and Taxon ID Tree 3 (Supplementary Materials) for Atomaria.

The DNA barcoding sequence from the mitochondrial gene COI-5P shows that this method can separated correctly 91% of morphologically distinctive species, including 75% for Cryptophagus and 94% for Atomaria. This reflect the highly intraspecific variability found within Cryptophagus compared with Atomaria.

Taxon ID tree 1 separated most genera well. All Atomaria are grouped togetherand Curelius is distinctly outside of Atomaria.All Cryptophagus are grouped togetherwith Myrmedophila and Henoticus at both ends. Myrmedophila seems to be a synonym of Cryptophagus, following Woodroffe & Coombs (1961) concept,according to the ID tree; however, the DNA sequence was incomplete, with COI-5P of 463 bp, much less than the usual 658 bp. Surprisingly, Salebius is incorporated within Cryptophagus though it is a distinct genus. Antherophagus, Pteryngium, Caenoscelis Henotiderus and Telmatophilus are well separated at the opposite end of the tree.

Taxon ID tree 2 shows a visually easily Nearctic identified species (C. mainensis)in the upper part, followed by three introduced species with toothed laterally prominent callosities without dorsal rim (C. distinguendus to C. scutellatus), followed by a species with callosities strongly produced anteriorly (C. tuberculosus), three species with elytra arcuate throughout and pubescence mostly suberect (from C. setulosus to C. valens), 6 mostly Palaearctic species with elytra mostly oblong and pubescence appressed, followed by very common Nearctic species (C. croceus and C. difficilis) and finally three common Holarctic northern transcontinental species (from C. lapponicus to C. bidentatus). The tree has shown (not illustrated) that many specimens identified as C. jakowlewi in Europe were in fact C. confertus. Of the 2 specimens seen by the first author and identified by Colin Johnson as C. jakowlewi, one was C. confertus and the other C. bidentatus. Esser (1994) showed clearly that all three species were distinct. The true C. jakowlewi has not been collected in North America and C. confertus remains a valid Holarctic species. Among species having two BIN numbers, we can mention C. tuberculosus, C. setulosus, C. dentatus and C. difficilis, all very common and widespread species.

Taxon ID tree 3 shows that most Atomaria species are well discriminated by the analysis, with all Anchicera (from A. mesomela to A. neomunda) being on one side and all Atomaria s. str. (A. nigrirostris to A. subdentata) being on the opposite side. We can see also some distinct groups. In Anchicera,some are more distinct as species with pronotum arcuate at middle and strong contrast black pattern at basal half of elytra (A. mesomela, A. distincta); antennae with A9-A10 transverse (from A. peltata to A. apicalis), within that group, dark species with pronotum borders entirely visible from dorsal view (A. nigritaria, A. peltata), and pronotum with sides subparallel at basal 0.5 (A. turgida, A. apicalis); the A. fuscata complex species group (from A. hudsonica to A. ocularia), antennae with A9-A10 subquadrate and pronotum border entirely visible from dorsal view (from A. kamtschatica to A. arcuaticollis) and pronotum with double-sided borders (from A. ornata to A. neomunda). In Atomaria s. str., we can see at first species with pronotum having strong basal groove (A. nigrirostris, A. impressicollis), with A9-A10 transverse and pronotum with sides not sinuate (A. tenebrosa, A. patens), A9 subquadrate and A10 transverse (A. pumilio, A. wollastoni), elytra with pubescence suberect (A. umbrina), pronotum subquadrate and elytra elongate with sides subparallel (A. linearis), A9 and A10 nearly subquadrate (A. vespertina, A. nigricollis) and A1 short and stout (from A. macra to A. subdentata). The tree also shows that the European concept of A. fuscata might be in fact three distinct species (one not included in this tree) that were synonymized probably by Johnson but this need to be verified. In the present publication, A. fuscata corresponds to the wide Johnson concept of the species (Johnson 1992). It would be very useful to look to all distinct Palearctic BIN species to see the external morphological characters that can separate them. It might prove that A. fuscata (= A. saginata) and A. ochracea are distinct species.

The analysis also shows that the A. wollastoni complex is in fact three species: A. wollastoni which is Palaearctic, A. pumilio and A. lineola, both being Nearctic, previously considered as synonyms of A. wollastoni. The tree also shows that A. pulchra was previously confused with A. vespertina. The analysis confirms the validity of 7 new species described in this publication (all with Pelletier as the author): A. arcuaticollis, A. neomunda, A. nigritaria, A. albertana, A. impressicollis, A. pinicola and A. puelloides. Three new species were discovered and described because of their distinct BINs (all with Pelletier as the author): A. ocularia, A. salicicola and A. calidaria, all from Alberta.