Devadatta ducatrix and D. kompieri - an appraisal

Until this year, the identification of Devadatta specimens in Vietnam seemed straightforward. There was one rather dull looking species called D. ducatrix and there was a slightly less dull species with a blue face called aptly D. cyanocephala. And because the situation was simple, I did not collect specimens of the dull species, which is a regular occurrence in hill and mountain forest in the north of Vietnam, where it inhabits shaded trickles.

This situation changed drastically after I last year explained the productive Yen Bai site to Toan and he visited it in the middle of May. He collected several Devadatta specimens and concluded that some represented an as yet undescribed taxon. This spring the new species was published in Zootaxa by Phan, Sasamoto and Hayashi. They did me the honor of naming the new species after me.

The differences between D. ducatrix and D. kompieri are slight and based on a comparison of very few specimens. But at the same time the complete set of differences should enable identification, although the authors stress that it is important to collaborate their findings with DNA analysis in the future, when the extent of the populations is more clear and more specimens are available.

Obviously I was keen to see this species carrying my name for myself and in the middle of May headed to Yen Bai Province. This year, I had decided, I would collect Devadatta specimens to compare. The first time I ran into the species was on May 14 still east of Nghia Lo on a mountain pass and almost 80km from Tu Le, where the species had been found by Toan. Later that day I also had them in that area. What is more, over the spring and early summer I noted the species regularly in Yen Bai, but also caught a male of what I concluded was D. kompieri near Lao Cai along the lower part of the road up to Sa Pa. In fact, I found few specimens of D. ducatrix this year, for instance in Thanh Hoa, Lang Son and Cao Bang Provinces.

I used a combination of characters from Phan et al. to identify the specimens, but also noted that some were not consistent. I caught for instance a specimen in which the number of cells between the first anal vein and the trailing edge of the hindwing (should be 2 in D. kompieri and 3 or more in D. ducatrix) was 2 in one wing and 4 in the other. Sadly it afterwards escaped when I tried to administer some moisture to it to keep it alive.

Today is a very rainy day in Vietnam and anyway, the season for Devadatta is over, so today I had a closer look at all specimens collected this season. I have 9 usable D. kompieri males, 1 D. ducatrix male and one female of both species (presumably), plus a scan of two presumed D. ducatrix males from Cao Bang.

Let's first summarize the differences described by Phan et al. for the two species. I paraphrase from the article, but slightly adjusted to exclude the references to D. yokoii.

(1) D. kompieri is smaller (Hw 31–34; abdomen including appendages 32–37) than the more robust D. ducatrix (Hw 33; abdomen including appendages 39 mm in our specimen, and each 40 and 43.3 mm in holotype (Lieftinck 1969)). (2) Wings of D. kompieri, compared with those of D. ducatrix, are proportionally shorter and have a less rounded apex and a narrower dark marking at tip. (3) The pterostigma is proportionally longer and narrower in D. kompieri than in D. ducatrix. (4) The venation of D. kompieri is less dense, especially in the space between first anal vein and posterior wing margin in the Hw, which consists of two cell rows at most, whereas there are two to three rows in D. ducatrix. (5) Superior appendages of D. kompieri are slimmer, while those of D. ducatrix are broader). 

The authors compared 4 specimens of the new D. kompieri with a single male D. devadatta also caught at Tu Le. Based on the extent of the dark marking at the tip of the wings and the shape of the pterostigma all my specimens from Yen Bai and the specimen from Lao Cai seem to be D. kompieri, whereas the one male from Thanh Hoa should on the basis of these characters be D. ducatrix, as should the two males I scanned from Cao Bang. If we look closer at the 9 D. kompieri specimens, length of abdomen (incl appendages) ranges from 36 to 41.5 mm (this is the specimen from Lao Cai, largest for Yen Bai is 41mm (2 specimens). This overlaps with both the one specimen identified by Phan as D. ducatrix and the size of one of the types from Lieftinck. My own D. ducatrix specimen has 38mm. Therefore it is questionable whether the size criterion is useful.

Hindwing length in my specimens ranged 32-38mm, again including the measurements of Phan's D. ducatrix specimen and also that of my own (34mm). This is logical, larger specimens will have larger wings generally. It illustrates that these measurements are of limited value to separate the species.

Relative hindwing length compared to abdomen length ranges 0.86 - 0.93 in D. kompieri (N=9) and 0.89 - 0.96 in D. devadatta (N=3). Although sample size is very limited, this may indicate that on average D. devadatta has slightly longer wings relatively. Nevertheless, given the overlap and small sample size, it remains to be seen whether in certain cases this may be of value.

In the pterostigma of the D. kompieri specimens (in 8 out of 9) the length of the pterostigma that is confluent with the costa is less than half that of the total length. In the D. ducatrix specimens it is slightly more than half. And the number of cells completely covered by the pterostigma is 4 in D. kompieri (one specimen 5), whereas in D. ducatrix it is 5 (one specimen 4). Therefore the shape and size of the pterostigmata in combination with the denser venation of D. ducatrix may be an indicator of specific identity.

However, the number of cell town between the first anal vein and the trailing edge of the hindwing appears to be of limited value and even varies within specimens between left and right. Sample size is limited, but number of cell rows (the maximum in the field) ranges 2-3 in D. kompieri (with only 2 specimens with maximum of 2 in both wings) and 3-4 in D. ducatrix (but one specimen with only 2 in one wing).

Last, I postulate that there is no consistent difference in the width of the apical part of the the superior appendages, whereas the inferiors are rather variable in outline in dorsal view. I consider the appendages of no value to support identification, but it must be stressed this is based on a single male D. ducatrix being no different from 9 males of D. kompieri.

In addition, although not given as an ID characteristic, the slightly amber colour of the wings in D. kompieri when mature also occurs in D. ducatrix when mature.

To summarize for the male specimens, apart from the size and shape of the apical spot on the wings, there is little to consistently support that this is a valid indicator of specific identity. Size is of no value, and shape and relative length of the wings either extremely limited or also of no value. It does seem that the density of wing venation may have some very limited usefulness, although D. ducatrix with only 2 cells between anal vein and trailing edge are to be expected, given that one of the 3 specimens examined has only 2 cells in one wing. The sample of D. kompieri does not contain any specimen with an abundance of 3 row cells or with 4 rows (although escaped presumed D. kompieri specimen had one wing with 4). In addition, the shape of the pterostigma also appears almost consistently to be different. Shape of the appendages appears not to be of value for identification.

When comparing the single females of both species (identified on the basis of extent of apical spot and association with males, although not in tandem) reveals that abdomen length of D. kompieri is 35.5mm compared to 32 for D. ducatrix. Hindwing is 34 in D. kompieri and 32.5 in D. ducatrix, thus supporting the conclusion that D. ducatrix is not consistently more robust than D. kompieri. Both specimens have 2 cells maximum between the first anal vein and the trailing edge, but D. kompieri has only 3 cells covered by the pterostigma against 4 in D. ducatrix, which is in line with the males. Whether this is consistent remains to be seen, given that the comparison is based on single females.

However, the cerci of the female D. kompieri are much shorter and more like an equilateral triangle than that of D. ducatrix. In D. kompieri to ovipositor extends beyond the tip of the cerci, in D. ducatrix this is reversed.

So, not all hope is lost. It is possible that D. kompieri and D. ducatrix are different, but clearly they are very similar species or subspecies. Differences in wing venation especially are of limited value when so slight as is the case here. Most differences appear clinal. Whether the differences are the product of peculiarities in geographical populations or truly represent separate species may require more research and, as the authors rightfully suggest, collaboration by DNA analysis. If it can be shown that the structural differences between the females hold up in larger samples this will support the case for two different species. 

Devadatta kompieri male from Lao Cai. The largest specimen and somewhat in between the two species as far as the shape of the apical spot in the fore wing goes. Number of cell rows and shape of pterostigma consistent with D. kompieri.

Same specimen from Lao Cai in lateral view

Male Devadatta kompieri from Yen Bai. Long pterostigma with few covered cells, two rows of cells in hindwing between first anal vein and trailing edge, limited apical spots.

Same male in lateral view

Rather typical Devadatta ducatrix male from Thanh Hoa. Note extensive apical spots, particularly in hindwing, more square pterostigma with 5 covered cells and 3-4 cell rows in hind wing between anal vein and trailing edge

Same male in lateral view

Two female Devadatta. Presumed D. ducatrix top and presumed D. kompieri bottom. Pointed face of latter due to opened mouth.

S8-10 of presumed female D. ducatrix. Note slender and long cerci

S8-10 of presumed female D. kompieri. Note shorter and stubby cerci.

Finally, two males of D. ducatrix from Cao Bang. Bottom male with limited extent of apical spots possibly result of very immature state, but specific identity not entirely sure (if we assume there are indeed two valid species). What is of interest here is that this specimen has 2 cell rows between first anal vein and trailing edge in right hindwing, but 3 in the left. In escaped presumed D. kompieri even more extreme difference observed of 2 in one wing and 4 in other.