[Odonata-l] Why report hindwing length and not forewing lengthalso?
Nick and Ailsa Donnelly
tdonelly at binghamton.edu
Mon Mar 19 14:51:58 PDT 2007
There are differing aims in data gathering. One can gather data for
internal purposes, that is, for studies involving one's own measurements
where the important aspect is that populations of A are measured the same as
for B. But if two different persons are measuring their own subpopulations
for a combined study, there is a real possibility that they will do the
measurements differently, and degrade the results. I got into this problem
with Ken Tennessen with our Macromia study. I measured over a thousand
specimens and he did several hundred more. We fretted that our measurements
might not be made the same way, so we exchanged detailed illustrations
showing how we did them. I think our data meshed pretty well, but we fussed
over this problem.
A great deal of measured data is not useful for general comparison because
there are probably built-in biases to the way in which the measurements were
made.
Why measure at all? It turns out that measurements may be useful for a lot
more than raw taxonomy. The tight correlation of abdomen length and hind
wing length (R = 0.99+) in Fijian damselflies told me that they were
consistently engineered for vigorous flight. I found an insular subspecies
that also regressed very well, but with heavier, and shorter abdomens, and a
different regression! Look what isolation did for these guys.
I later studied a large isostictid damselfly from New Britain that had a
very poor correlation. Though I never saw it alive, it seems that perhaps
it is a much poorer (or different) flier, like some of the pseudostigmatids
("helicopters") of the New World tropics.
Like so much scientific enquiry, we look and may find nothing. But then we
notice something of interest and are glad we looked. Is this efficient?
Certainly not, but it can be very interesting.
-----Original Message-----
From: aardila at uoguelph.ca [mailto:aardila at uoguelph.ca]
Sent: Monday, March 19, 2007 1:18 PM
To: 'Odonata-l'; Ola Fincke; Nick and Ailsa Donnelly; 'Mike May'
Cc: John and Sue Gregoire <khmo at att.net>
Subject: Re: [Odonata-l] Why report hindwing length and not forewing
lengthalso?
Hi everyone,
I have really enjoyed the discussion so far on this subject. I think that
there is always going to be experimental error (EE) in this type of
measurements even if there is a detailed protocol. Some initial questions to
me are: is the current error statistically significant?
Do mean values of individuals from a given species measured by different
people vary significantly due to EE? I think that from what it has been
discussed EE is fairly small (a few milimeters), therefore this may not be
significant and as Mike hinted it would take too much time and unnecessary
effort to make detailed (not rapid) measurements of every individual. The
mean values seem to be fairly close to the mean values that can be
calculated from NWM and WM. Therefore, I think that it may not be necessary
to develop a detailed protocol to measure wings sizes and body sizes unless
it is required for a detailed study in which different people are making
measurements with a specific purpose in mind. The current methods seem to be
working well.
I think that some of the most important question are, what is the biological
significance of all these raw data? Does it matter biologically if two
people measure the same individual and get measurements that differ in 2mm?
Does it matter if we think that a given species have wings mean sizes of 20
or 18mm? I think that it may not matter. I think it may be more important to
look at different size ranges and go from there. That is for example, to
look at Odonata species that have wings mean values of less that 20mm,
20-40mm, and more than 40 mm, and see if there is anything that these
species have in common biologically-translating the raw data to a biological
meaning.
Then, what about intraspecific variation? That is, body size variation
within species. There seem to be a lot of factors that influence body size
in insects. I am only starting to get into this literature, but I have read
for example that time, temperature and food are usually important factors
that account for this variation. Also, latitud and what it has been
mentioned in this discussion seasonal variation. I think that seasonal
variation encompasses so many factors that there is room for multiple
studies, if not done before, as to what are the most important factors that
influence body size variation in Odonata.
One of them is sex, as Mike mentioned, females tend to be larger than males
in Damselflies and that is reported in his book with the range values for
body size. But what about other factors such as food?, for example, I wonder
if it matters what nymphs eat, Does this have an effect on adult body size?
Does it matter if there is absence or presence of some nutrients in the soil
or pond where they live?, the same questions apply for temperature and other
factors.
So, I think that the discrepancy between measurements may be explained and
be more important from the biological perspective as oppposed to how
measurement are done. The EE seems fairly unsignificant. Also, the mean
values don't seem to differ significantly. If they do vary significantly,
then we may wonder if the person making the measurements made a mistake, or
more interestingly, maybe there is something really exiting going on at that
location with those dragon/damsel flies.
Alex
Quoting Ola Fincke <fincke at ou.edu>:
> All my measurements are on live insects, so I just measure from the
> tip to the base of the wing (i.e. end of the wing itself, not the
> structure into which it inserts).
> To do the abdomen, I just tap it straight and measure. Obviously
> there's more error measurement associated with that measure, than the
> wings.
> I have also done wet weights, to get correlations among wing, abdomen
> and wet weight (all are highly correlated). Brad Anholt has done a
> lot of the latter.
>
> Ola
>
> Ola Fincke
> Dept. of Zoology
> University of Oklahoma
> Norman, OK 73019
> Tel: 405-325-5514
> Fax: 405-325-6202
>
>
>
>
> On Mar 17, 2007, at 1:48 PM, Nick and Ailsa Donnelly wrote:
>
>> This interests me a lot, because I have been measuring Odonata for
>> many years.
>>
>> One problem with wing measurement is that it is not certain where
>> the base of the wing is located. In my recent Orthemis study I
>> have been taking the notch at the base of the costa that seems to
>> correspond to the major point of flexure of the wing. This is
>> pretty difficult to observe if one is using a millimeter scale on a
>> specimen held beneath the microscope, but it can be done.
>> Tillyard more or less shows where it is, but he does not figure it
>> worth a damn. My technique is to scan the wing and scale
>> together, and then to transpose the scale image on to the wing in
>> Photoshop, rotating and translating the scale to make the
>> measurement. In this way I can get about 0.1 mm repeatability.
>> This is great for my own exercises, in which I wish to show
>> comparisons with other specimens measured also by me. But how do
>> I compare my results with those of others? The simple answer is
>> that I cannot do this, because I have no idea how others measure
>> wings. So, how do you do it, and from where?
>>
>> And abdominal measurment - don't even talk about that. Even when
>> the abdomen is still in one piece, and it is straight, not curved,
>> there is the problem of shrinkage during preservation. This
>> measurement is vastly less certain than wing measurement, and I use
>> the hind wing as the only more or less reliable surrogate for
>> size. I still am not quite certain what all this is
>> accomplishing. I found in the Orthemis study one specimen for
>> which I thought I had made a serious mistake - it couldn't really
>> be shaped like that, with such short wings and a long abdomen. I
>> pulled the specimen out and looked at it again- It really was
>> shaped like that - long abdomen and short wings. No mistake at
>> all. But what was the meaning of its reported size?
>>
>> What I would like to see done is weight measurements. I once found
>> a species of Nesobasis (damselfly) for which the males tended to
>> be much larger than the females. I estimated from dimensions
>> alone that the males must weight at least twice what the females
>> weighed, which I think is remarkable. But a protocol for weight
>> measurement would be very difficult and would have to involve
>> taking a very portable and accurate scale into the field.
>>
>> A digression. Years ago I was in my office when the friendly
>> Fishers sales rep wandered in. "What do you have that is new and
>> interesting?". I asked. "Why we have just introduced portable,
>> digital, battery powered scales", he said with considerable pride.
>> "I'll bet the drug dealers are scarfing them up", I replied. He
>> promptly turned an interesting shade of gray, never having
>> considered this marketing possibility.
>>
>> From: odonata-l-bounces at listhost.ups.edu [mailto:odonata-l-
>> bounces at listhost.ups.edu] On Behalf Of Ola Fincke
>> Sent: Thursday, March 15, 2007 4:16 PM
>> To: Mike May
>> Cc: Odonata-l
>> Subject: Re: [Odonata-l] Why report hindwing length and not
>> forewing length also?
>>
>> In many species, it's well-known that size declines over the
>> season, which is one reason that those of us interested in the
>> effects of size on reproductive success take that into account
>> (e.g. Fincke & Hadrys 2001; see also work by Dave Thompson and
>> others, cited in the former).
>>
>> There must be a LOT of geographic variation in size; certainly
>> Ischnura verticalis in MI is much larger than it is in OK. Even
>> within species in similar areas, size can vary quite a across
>> populations. I am collecting data on this within Enallagma.
>>
>> Ola Fincke
>> Dept. of Zoology
>> University of Oklahoma
>> Norman, OK 73019
>> Tel: 405-325-5514
>> Fax: 405-325-6202
>>
>>
>>
>>
>> On Mar 15, 2007, at 1:52 PM, Mike May wrote:
>>
>>> Dennis, et al.,
>>> I have to agree that a much more useful set of measurements,
>>> regardless of which body parts are selected, would be mean +/-
>>> standard deviation (range). Likewise it would be great to include
>>> much more precise distribution data (e.g., county rather than
>>> state level - hooray again for Nick Donnelly's dot maps) and
>>> seasonal data (e.g., flight season at several different latitudes
>>> or within regions of similar climate). The reason for not doing
>>> so in NWM and W&M is simple, and the same in all these cases -
>>> time and effort. To a fairly accurate approximation, one can
>>> (read, "Mike can") riffle through the hundreds of specimens of
>>> Enallagma civile at FSCA, pick out several of the largest and
>>> smallest from different regions, spread them out and compare by
>>> eye, select and measure 3-4 at each end of the size range, and
>>> write down the maximum and minimum results. That's exactly what
>>> I did, and, although tedious enough, it's a lot quicker than
>>> selecting a random sample large enough to be statistically
>>> relevant (~30), measuring (with a mm ruler, sometimes viewed
>>> through a scope) and recording all of them, and calculating the
>>> appropriate statistics. That doesn't make the former technique
>>> great, especially if one or two atypically large or small
>>> individuals are incuded, it's just easier. There's also no doubt
>>> that that human error plays a role and is likely to lead to a more
>>> misleading outcome in assessing range than aggregated statistics.
>>> As an aside, total length and abdominal length are probably much
>>> more subject to error than is wing length because of the tendency
>>> of heads to fall off, abdomens to curl, etc.
>>> Anyway, I think a genuinely worthwhile project would be to
>>> accumulate and publish careful size measurements of as many
>>> odonate species as possible, based on statistically meaningful and
>>> geographically diverse samples. This just wasn't feasible in
>>> doing the manuals, but it would be very useful in itself and
>>> could lead to some interesting questions. E.g., does size vary
>>> geographically or seasonally, and is the variation consistent
>>> across species? This has been shown to be true for several
>>> species, as Dennis notes, but the answer isn't known for most. If
>>> such patterns do exist, then you can ask why. If anyone is
>>> interested in undertaking such a project, I'd be glad to give
>>> them the benefit of my experience, for what it's worth.
>>> Mike May
>>> ----- Original Message -----
>>> From: Dennis Paulson
>>> To: Odonata-l
>>> Sent: Wednesday, March 14, 2007 12:17 PM
>>> Subject: Re: [Odonata-l] Why report hindwing length and not
>>> forewing length also?
>>>
>>> What is sometimes frustrating is that guidebooks such as these
>>> list measurement extremes without giving any idea of means or
>>> explaining the significance of the extremes, and the handbooks
>>> Alex mentioned are among the best examples of this.
>>>
>>> For example, I opened Needham, Westfall & May to a page at random
>>> and found Erpetogomphus bothrops. Length 38-51 mm, abdomen 26-39,
>>> hindwing 21-31. This is a common species in Sonora, which I've
>>> visited several times in recent years, and I always thought the
>>> individuals I saw and collected didn't vary anything like that. So
>>> I just measured the hind wings of all my specimens (from
>>> throughout the range of the species) and found this:
>>>
>>> males - range 23-26.5 mm, mean 24.9 (n=15)
>>>
>>> Eleven were with n 1 mm of the mean, in the range 24-26 mm.
>>>
>>> females - range 26.5-29 mm, mean 27.5 (n=6)
>>>
>>> Five were within 1 mm of the mean, in the range 26.5-28.5 mm.
>>>
>>> Thus in both sexes, the great majority of individuals measured
>>> within a 2-mm range, and that range defines the size of each sex
>>> quite well. Combining the sexes would justify listing the range in
>>> "usual" hindwing length of 24-28.5 mm. Total length and abdomen
>>> length don't seem to vary any more than hindwing length.
>>>
>>> The entire series of 21 individuals ranged from 23-29 mm. Rosser
>>> Garrison, in his 1994 revision of the genus, examined 91
>>> specimens. He listed the variation in hindwing length as 27-31 mm,
>>> which makes me wonder if we measure wings differently. In fact
>>> we must, because some of my specimens were among those he
>>> measured. Both of u show a considerably smaller range than that
>>> listed in NWM, but between my measurements and Garrison's
>>> measurements, we're approaching the range in NWM, so part of the
>>> great variation listed in that book could be explained by
>>> different measurement techniques and/or acquiring the
>>> measurements from all published data. Unfortunately, ranges like
>>> that don't give one a very good idea of how big the dragonfly
>>> is. Perhaps Mike May can tell us how those figures were
>>> obtained. I emphasize that I'm not criticizing NWM, as I'm sure
>>> the authors of that book did the best they could in acquiring
>>> these figures.
>>>
>>> I do know, from much research on bird size, that guide book
>>> writers sometimes get it wrong, and the incorrect figures have a
>>> life of their own, copied from book to book. I found total length
>>> measurements off by as much as 15% when analyzing measurements in
>>> several bird guides, at times a larger species listed as smaller
>>> than a smaller species. How many bird-b ok writers nowadays
>>> measure birds, live or dead?
>>>
>>> In many taxonomic papers, the author measures a series of
>>> specimens and gives ranges and means, but for many species there
>>> are no such papers, so writers of guide books have a difficult
>>> time getting measurements, with two options. Either they have
>>> access to a collection and spend long hours measuring specimens,
>>> or they take the figures from other publications, the research
>>> literature and/or other guide books. One of my colleagues recently
>>> confessed to me that when he listed measurements for a book on
>>> dragonflies, he just used the midpoint of the measurements in NWM
>>> and Westfall & May, which seems as good a way of doing it as any.
>>> I think having a single measurement in a guide book makes it
>>> easier to picture the size of the animal than when there is a
>>> rather substantial range.
>>>
>>> Some species vary in size over the season. Pachydiplax longipennis
>>> t at emerge in summer in southern Florida have hind wings about
>>> 3 mm shorter than those that emerge in the winter, and the size
>>> varies seasonally at least as far north as the US Gulf states. I
>>> am not aware if seasonal variation has been established for other
>>> species. Just from eyeballing it, it seems to me that some
>>> species vary more in size than others, but perhaps that just
>>> depends on how many of them you see. I was just examining two
>>> Sympetrum costiferum, one of which seemed no more than 2/3 the
>>> bulk of the other, and two Enallagma carunculatum, one of which
>>> was surely only half the bulk of the other and about 2/3 its
>>> length. These are unusual extremes, of course. When I look at my
>>> collection or at odonates in the field, individuals of a species
>>> mostly look around the same size.
>>>
>>> Another source of variation in odonates is temperature.
>>> Damselflies in Mexico and Central America tend to be larger at
>>> higher elevations, presumably something about temperatu e vis-a-
>>> vis development time of the larvae. I don't know if there is
>>> latitudinal variation as well, but there should be. Nick Donnelly
>>> called attention to the very small size of some damselflies in
>>> eastern Texas, and I have seen these same populations of
>>> impressively tiny creatures. I don't know whether these
>>> populations are genetically distinct, or does some environmental
>>> factor of the area have them emerging at a smaller size? I
>>> collected tiny Perithemis mooma in Yucatan, as small as or smaller
>>> than Nannothemis bella.
>>>
>>> There are probably many more examples of interesting size
>>> variation in odonates, and it would be good to know more about this.
>>>
>>> Dennis Paulson
>>>
>>> On Mar 14, 2007, at 6:58 AM, Mike May wrote:
>>>
>>>> Certainly part of the reason is tradition. Once a large body of
>>>> data is available for a particular measurement, the value of that
>>>> measurement for comparative study is enhanced. I suspect, also, that
the hindwing
>>>> may have
>>>> been selected originally because, when a dragonfly is held in the
>>>> hand with
>>>> the wings pressed together above the back, the hindwing is positioned
>>>> outside the forewing. Although the forces produced by fore- and
>>>> hindwings in
>>>> flight are certainly somewhat different, I feel pretty confident
>>>> that had nothing to do with the decision.
>>>>
>>>> The forewings are nearly always slightly longer than the hindwings. My
>>>> impression is that the proportionate difference does not vary
>>>> greatly with
>>>> size, taxon or sex, but I don't know whether this has been
investigated
>>>> rigorously. Females do typically have longer wings than males in
>>>> comparison
>>>> to other measures of body size such as total length or thoracic mass
>>>> (probably because abdominal mass becomes greater in mature
>>>> females because
>>>> of the mass of eggs), but I don't know that the differences
>>>> between fore- < SPAN>
>>>> and hindwings is affected by sex. Maybe Roy Beckemeyer can
>>>> illuminate that
>>>> question a bit more.
>>>>
>>>> Mike May
>>>>
>>>> ----- Original Message -----
>>>> From: <aardila at uoguelph.ca>
>>>> To: "Odonata-l" <odonata-l at listhost.ups.edu>
>>>> Sent: Wednesday, March 14, 2007 4:28 AM
>>>> Subject: [Odonata-l] Why report hindwing length and not forewing
>>>> length also?
>>>>
>>>>
>>>>> Hello everyone,
>>>>>
>>>>> I have the books "Dragonflies of North America"(Needham, Westfall
>>>>> ,and May, 2000) and "Damselflies of North America" (Westfall and
>>>>> May, 1996). These books are excellent. They report the body size
>>>>> lenght, abdomen lenght, and hindwing length for every Odonata
>>>>> species in North America.
>>>>>
>>>>> However, I don't understand why the size of the forewings is not
>>>>> reported. Is this by convention? What is the history behind this?
>>>>> Why were the hindwings and not the forewings chosen? Is there a
>>>>> statistically and biologically insignificant difference between
>>>>> the forewings and the hindwings within species (and sex) in
>>>>> Odonates? Is this selection based on the role of the hindwings in
>>>>> flight? I would appreciate any guidance that could help me find
>>>>> an answer to these questions. Thank you.
>>>>>
>>>>> Alex
>>>>>
>>>>> --
>>>>> Alex Ardila-Garcia
>>>
>>> -----
>>> Dennis Paulson
>>> 1724 NE 98 St.
>>> Seattle, WA 98115
>>> 206-528-1382
>>> dennispaulson at comcast.net
>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Odonata-l mailing list
>>> Odonata-l at listhost.ups.edu
>>> https://mailweb.ups.edu/mailman/listinfo/odonata-l
>>> _______________________________________________
>>> Odonata-l mailing list
>>> Odonata-l at listhost.ups.edu
>>> https://mailweb.ups.edu/mailman/listinfo/odonata-l
>>
More information about the Odonata-l
mailing list