Charles S. Henry.
Temperature-Induced changes in the calls of the Green Lacewing, Chrysoperla plorabumla (Neuroptera: Chrysopidae).
Psyche 90:343-360, 1983.

Full text (searchable PDF, 1224K)
Durable link: http://psyche.entclub.org/90/90-343.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

TEMPERATURE-INDUCED CHANGES IN THE CALLS OF THE GREEN LACEWING,
CHRYSOPERLA PLORABUNDA (NEUROPTERA:
CHRYSOPIDAE)*
BY CHARLES S. HENRY
The Biological Sciences Group
Box U-43, University of Connecticut
Storrs, Connecticut 06268
Animals communicate acoustically in many different ways and for many different purposes; in fact, a vast literature exists on the subject, and comprehensive efforts to summarize current knowledge have not been attempted since the early 1960's (e.g., Lanyon and Tavolga 1960, Busnel 1963). Insects are especially rich in singing or noise-making species, and it seems likely that every insect order will eventually prove to include acoustically active taxa. Neuroptera has long been regarded as a "silent" order of primitive, behaviorally simple insects. However, recent work suggests that many, if not most, species of the large neuropteran family Chrysopidae are characterized by complex courtship displays accompanied by a specialized form of acoustical signalling (Smith 1922, Toschi 1965, Ickert 1968. Henry 1979 and 1983b). Such lacewing calls or songs are not acoustical in the traditional sense, but instead consist of species-specific substrate-borne vibrations produced by vigorous, stereotyped jerking motions of the insects' abdomens: a phenome- non known as tremulation in other insects (Morris 1980; Henry 1980a, c), and found also in the ancestral neuropteroid taxon Megaloptera (Rupprecht 1975). Calling behavior is most elaborate in Chrysoperla Steinmann; conspecific males and females of species within that genus cannot mate without first reciprocally exchanging similar or identical vibrational signals in a prolonged courtship duet (Henry 1980b. 1983b). The best studied species in this regard is Chrysoperla plorabunda (Fitch), a common North American form that for some years was considered synonymous with the morpho- logically identical Eurasian species Ch. cornea (Stephens) (Tjeder 1960. Henry 1983a). Both sexes of this species, when sexually å´Manuscrip received by the editor AU~US! 2, 1983



================================================================================

344 Psyche [VOI. 85
receptive, produce brief, repetitive volleys of low-frequency abdom- inal vibration at (approximately) one-second intervals for up to several minutes at a time (Henry 1979, 1980~); in heterosexual duets, the two partners alternately and repeatedly exchange single volleys until copulation is achieved. Each volley in Ch. plorabunda is characterized by a smoothly increasing and then decreasing ampli- tude envelope and by pronounced frequency modulation (see Fig. I): the rate of abdominal vibration gradually declines to less than half its initial value during the course of the volley (Henry 1980~).
Green lacewings of the genus Chrysoperla are unusual among acoustical animals in that females are just as likely to sing as males (Henry 1983b). In contrast, males alone (or principally) sing and call conspecific females to them in most species of the best studied animal taxa like birds (Catchpole 1982), lizards (Frankenberg 1982), frogs (Littlejohn 1977), katydids (Dumortier 1963), homopterans (Ossiannilsson 1949, Young 1980), and crickets (Walker 1962, Dumortier 1963). Such unilateral male signalling has been inter- preted in most acoustical animals as indicative of sexual selection operating on the higher variance in reproductive success of males (Halliday 1978, Otte 1974, Alexander 1975). On the other hand, the unusual presence of calling behavior in both sexes of Chrysoperla species may stem from the proven importance of their vibrational signals in the reproductive isolation of closely related species. For example, neither of the sibling, interfertile species Ch. plorabunda and Ch. dowesi (Banks) will respond to the song of the other, thus precluding courtship duets and preventing interspecific mating under natural conditions (Henry 1983b). If such signals are to be effective as reproductive isolating mechanisms, however, they must remain unambiguous to recipients over the wide range of tempera- ture typically experienced by lacewings in the field. Abundant doc- umentation exists of gross alteration in chirp rate, wing-stroke frequency, pulse or chirp duration, or call notes by temperature changes in many taxonomically disparate insect groups (Brooks 1882, Hayward 190 1, Alexander 1956, Walker 1962, Dumortier 1963, Shaw 1968, Booij 1 982), and similar temperature-related changes should be characteristic of lacewing songs (see Henry 1982b for preliminary work on Chrysopa oculata Say). One would also predict that the calls of the two sexes of a given species of Chryso- perla should vary in a closely parallel fashion over a wide range of



================================================================================

I 9831 Henry - Chrysoperla plorabunda 345 temperatures, since mutual recognition of highly specific call fea- tures is so important to the reproductive success of both members of each courting pair. Here, I report on the effects of temperature change on the principal parameters of the calls of individual male and female Chrysoperla plorabunda from North America. This paper also contains the first complete description of the frequency structure of the abdominal volleys of that species, together with an experimental analysis of the effects of abdominal mass on frequency characteristics. Regressions of call parameters against temperature in Ch. plorabunda are compared with those described for other singing insects, in an attempt to identify any unifying principles. A breeding colony of Chrysoperla plorabunda was started in the fall of 1982 from seven males and ten females collected in a field of senescent goldenrod (Solidago spp.) at Storrs, Connecticut. Subse- quently, adults were maintained on a WheastTM/sucrose diet while larvae were fed ether-killed Drosophila spp. (see Henry 1979, 1983a for details). Males for experimentation were drawn from second- generation laboratory stock, while females were third-generation insects; five unmated individuals of each sex were acoustically moni- tored at various temperatures between 19.5OC and 29.8OC. For each of five call characteristics of interest, I analyzed an average of 40 volleys of abdominal vibration per individual, delivered at three to six different temperatures between the extremes mentioned above; in no case was a regression line for an individual based upon fewer than 18 volleys. Temperature was monitored within 25 cm of the calling insects, and was controlled by heating and cooling an entire 120 cubic meter room.
Lacewings were induced to call either by playing back to them cassette tape recordings of conspecific signals or by simulating such signals by means of a sweeping audio frequency generator (Tek- tronixTM FG 507) gated by a physiological stimulator (Harvard lM 340). Patterns of abdominal vibration were detected and analyzed with techniques and equipment described in other papers (Henry 1980a, 1982b). Details of call parameters were obtained from Pola- roid photographs of oscilloscope tracings, using conventional overlay methods.
Several individuals of Ch. plorabunda from a breeding colony of different geographical origin were selected for use in experiments



================================================================================

346 Psyche [VOI. 85
testing the effect of abdominal weight on vibrational freq uency. This stock was bred from five males and five females collected in the sagebrush country of southwestern Idaho on 24 May, 1983 by Dr. James Johnson (University of Idaho). I tested one field-caught male, two first-generation laboratory males, and one first-gener- ation laboratory female for the frequency characteristics of their calls at two different temperatures; after ascertaining that a 11 were essentially identical to one another and to Connecticut plorabunda in those characteristics, 1 weighed each individual to the nearest tenth of a milligram on a MettlerTM H6T or SartoriusTM 12 12 MP balance and added weight to the middle of their abdomens, using water-based Liquid PaperTM. Mass-loaded specimens were then re- weighed and tested again for the vibrational frequencies of their calls at temperatures in the 25-30å¡ range. All such frequency values were also adjusted to 27OC using linear regression C in Fig. 3. Curve-fitting of paired variables to linear, exponent ial, or logarithmic functions employed a program designed for a H ewlett- PackardTM HP-25 pocket calculator. Any reference in the Results or Figure 1.
Detailed fragment of the call of Ch. plorahunda, re-drawn from an oscilloscope tracing, showing principal parameters A-E defined in text. A = initial volley frequency. B = median volley frequency, C = terminal volley frequency, D = volley repetition rate, E = volley duration.



================================================================================

19831 Henry - Chrysoperla plorabunda 347 Discussion sections to "significant differences" indicates that the means of two normally distributed samples were demonstrated to differ from one another by a 2-tailed t-test using confidence limits of 95% or better. Values following a +/- sign are one standard deviation of the mean.
I measured five different characteristics of the calls of Chyso- perla plorabunda males and females. These are listed and defined below and illustrated in Fig. 1
Initial volleyfrequency: The cycling rate per second of the first eight strokes of the abdomen, at the start of a volley of abdominal vibration.
Median volleyfrequency:
The cycling rate per second of the
eight abdominal strokes following the initial period defined above.
Terminal volley frequency:
The cycling rate per second of
the abdominal strokes that remain in a volley after A and B have been deleted.
Volley repetition rate: The rate per minute at which a calling lacewing produces volleys of abdominal vibration. Volle-Y duration:
The length of time in seconds required for completion of a volley of abdominal vibration. Frequency of abdominal vibration versus temperature is tabu- lated for all males and females in Table 1 and expressed as linear regressions for each sex in Figure 2. Since vibration frequency decreases during the course of a plorabunda volley, it is necessary to subdivide each volley into the three portions A, B, and C, defined above; frequencies measured for those portions are plotted sepa- rately on the graph. Other call parameters like volley repetition rate and volley duration are similarly plotted separately for each sex on the same figure, using different units on the y-axis. All data are combined for both sexes in Figure 3, which also displays the range of variation in the calculated regression lines for all 10 individuals. Several obvious features emerge from these tables and plots. First of all, it appears (Table 1) that males and females differ from each



================================================================================

Volley Repe- Volley Number of
Vibration Frequency. strokes/ sec
tition Rate, Duration, Abdominal
Temperature Sex Initial (A) Median (B) Terminal (C) volleys/min (D) msec (E) Strokes/Volley Male 71.55
? 3.41 (44)
Female 69.23
? 2.49 (63)
Male 86.19
? 3.60(37)
Female 84.40
? 2.91 (56)
Male 98.75
? 5.86 (47)
Male å± 12.35
? 5.52 (57)
Female 1 13.25
2 4.16(71)
Table I.
Principal characteristics of the calls of males and females of C7;n'wpei'hi plorahiimla. as measured at several n
different temperatures. Mean values and their standard deviations are tabulated. Sample sim are entered parenthetically: in each row. the same five males or five females produced the given number of measured volleys. Parameters A F. are depicted 5
-
00
in Fig. I and defined in the text.
(-x




================================================================================

19831 Henry - Chrysoperla plorabunda 349 other very little, at any temperature, in any major characteristics of their calls. Particularly coincident between males and females are the initial volley frequencies and the volley durations (Table 1 and Fig. 2A and E) of Ch. plorabunda calls. There is a consistent ten- dency for females to vibrate their abdomens in later portions of their volleys at somewhat lower frequencies than do males (35 cycles/ sec. vs. nearly 40 cps at 20å¡C see Fig. 2C) and for volleys to be produced rather more slowly by males than by females (about 66 volleys/ min vs. nearly 70/min at 29.5OC; same figure, D), but neither of these differences is statistically significant. Secondly, it can be seen from Figures 2 and 3 that the slopes for the linear regressions of frequency versus temperature differ radi- cally from one portion of a volley to another, gradually becoming less steep as the volley progresses. Thus, initial portions of volleys change frequency rapidly with temperature (slope = 4.27~; Fig. 3A), while terminal portions remain within a much narrower range of values over equivalent temperature extremes (slope = 0.960~). X- intercepts also differ significantly for each of the three regressions calculated from pooled frequency data on all ten individuals: inter- cepts of 3.60, - 1.27, and - 1 9.0g0 C respectively characterize initial, median, and terminal portions of volleys. Regression of another call parameter, volley repetition rate, against temperature produces a line with a slope of 3.25~ and an X-intercept of 9.14O C (Fig. 3D). Thirdly, Figure 3E demonstrates that volley duration decreases markedly with increasing ambient temperature, and that the mathe- matical relationship between the two can equally well be interpreted as linear or exponential. In theory, the total number of abdominal strokes per volley could remain constant as temperature varies, since the increased frequency of abdominal vibration at higher tem- peratures would automatically shorten volley duration. However, data in Table 1 suggest that higher temperatures induce a slight but significant reduction in the total number of abdominal strokes pro- duced during each volley, and that this phenomenon facilitates the volley-shortening process.
Finally, the table and figures all support the view that deviation is relatively slight between the temperature data for each of the var- ious call parameters and the linear regressions calculated from those data. The closest fit to a mathematical relationship is found in the initial frequency of abdominal vibration during a volley: for data



================================================================================

Psyche
[Vol. 85
1100
1000
900
m
-0
c
u
z
.-
- -
.-
E
z
0
h-
a
0:
3
n
400
Figure 2.
Regression lines. calculated from data in Table I, showing the effect of temperature on the five principal characteristics A E of the calls of five males (dashed lines) and five females (solid line) of Ch. pI~~ruhunc/u. The left-hand "frequency"axis applies to parameters A-C (strokes/sec) and to D (volleys/min), while the right-hand axis for "duration" applies only to E.
A. males:
females:
9. males:
females:
C. males:
females:
D, males:
females:
E, males:
females:




================================================================================

19831 Henry - Chrysoperla plorabunda 35 1 pooled from all ten individuals, over 95 percent of variance is explained by regression line A shown in Figure 3. Even for the worst case-terminal volley frequency-nearly two-thirds of raw data var- iance is compatible with the calculated linear regression (r2 = 0.65). And data for individuals are neither more nor less variable, on average, than pooled samples: for the five principal call features graphed in Figure 3, individual r2 values average 0.93, 0.84, 0.59, 0.79 and 0.76 and never fall below 0.40 for any single insect. Also shown in that figure is the close congruence of all individual lines for each of the same five call features but particularly for initial volley frequency, suggesting that such temperature relationships are con- sistent, repeatable, and predictable on an individual basis. Results of experiments manipulating abdominal mass in individ- ual lacewings are shown in Figure 4 and Table 2. Table 2 presents the raw temperature-frequency data taken from four insects, while Figure 4 shows how those data relate to the linear regressions gener- ated from the terminal volley frequencies of the ten unmodified individuals tested earlier. Converting the frequency measurements to their equivalent values at 27OC (Table 2) dramatically reveals how little those data are affected by mass-loading of the abdomen: in none of the four experimental animals was the terminal volley frequency altered significantly by the treatment. Weight increments (from painting) amounted to 10-27% of total body mass; however, abdominal weights were only 2.8 mg for the female and approxi- mately 2 mg for the three males, so increments to the mass of the vibrating structure itself in each insect actually ranged from 36 to 50%.
The results described above amply document the striking similar- ity of male and female calls of Chrpoperla plorabunda, consistent with the proven importance of such signalling behavior to the reproductive isolation of this species from several of its morphologi- cally identical siblings in the genus (Henry 1980b, 1983b). Females closely resemble males in every detail of volley structure and spac- ing, and those few differences that do exist in pooled samples tend to break down when the characteristics of individual insects are compared (Figs. 3 and 4). Also as predicted, males and females



================================================================================

Psyche [Vol. 85
Figure 3.
Total range of individual variation of temperature regression lines for the five major call characteristics A-E of Ch. plorahunda. The area above and below each straight line includes within it the linear regresssions for all 10 individuals tested, while the line itself summari7es temperature data for those 10 insects in a single regression equation. Axes and labels as in Fig. 2. A: y = 4.27~ - 15.37, r^ 0.95
B: y = 2.69~ + 3.41. r-^ 0.88
C: y = 0.96~ + 18.33. rz= 0.65
D: y = 3.25~ - 29.72, rz= 0.83
E: y = - 39x + 1620, r? = 0.83




================================================================================

19831 Henry - Chrysoperla plorabunda 353 change their various call characteristics in precisely parallel ways as temperature is altered, so that the sexes' calls remain indistinguish- able over the range of temperatures they would typically encounter. The only apparent exception to this conclusion concerns the slightly lower vibration frequencies characteristic of the terminal portions of the volleys of females at all temperatures (Fig. 2C). Such a differ- ence is expected if the terminal volley frequency of a lacewing's abdomen is determined by its inherent capacity to oscillate like a weight on a spring, since the demonstrably heavier abdomen of a female would resonate at a lower frequency than that of a male. However, artificially mass-loading the abdomens of several males and females produced no obvious downward deflections of their frequency characteristics, suggesting that neuro-muscular mecha- nisms actively "drive" the vibrating system for the entire duration of each volley (Table 2) and that resonance effects have relatively little influence on resultant frequencies. Also, as mentioned earlier, the male-female difference in terminal volley frequencies is considerably less impressive when the responses of individual insects are dissected from the pooled data (Fig. 41, thus raising the suspicion that it is an artifact of some sort.
Studies of other acoustical insects overwhelmingly support the existence of linear functions relating temperature to most song parameters that repeat over time, as best exemplified by and documented for rates of wing-stroking, chirping, and "rolling" in crickeLs (Alexander 1956, Walker 1962, Dumortier 1963, Prestwich and Walker 198 1) and a few katydids and homopterans (Dumortier 1963, Shaw 1968, Whitesell and Walker 1978, Booij 1982). Similarly, most of the temperature data reported here for Ch. plorabunda conform well to linear statistical models, although they are insufficiently detailed to discriminate linear from exponential interpretations. The least individual or pooled variance from the calculated regression is found for the abdominal vibration frequency of the first eight cycles of a volley, suggesting that this feature of the call is particularly crucial to unambiguous communication between the sexes; otherwise, it seems there would exist no need for such precision. In his 1962 paper on cricket song, Walker drew attention to the apparent convergence of many of his linear regressions on 4O C; that is, it seemed that for many different cricket species the chirp



================================================================================

Table 2. Effect on terminal volley frequency (C). in abdominal strokes/sec, of artificially mass-loading the abdomen of three males and one female of Ch. plora- bunda. Weight gains are tabulated in column 3: the last column lists n, the number of volleys measured. Equivalent frequencies (at 27OC) in column six were calculated from the slope of regression line C in Fig. 3. Superscript "x*' denotes lab-reared insects.
Weight. Temp.. Vibration Equivalent
mg.
OC. frequency.?SD at 27OC. n
Male 1 control 7.5
26.8 39.78 2 1.83 39.98 29
mass-loaded 9.0
30.0 47.08 k l,84 44.00 20
Male 1' control 6.7 27. I 44.49 k 2.44 44.51 20 mass-loaded 8.5 27.0 44.59 k 3.32 44.59 25 Male 2' control 7.2 27.3 45.98 ? 2.44 45.60 36 mass-loaded 8.5 26.8 44.92 k 2.80 45.15 36 Female I ' control 9.6 27.3 43.47 ? 2.31 43.20 28 mass-loaded 10.6 27.0 43. I4 k 2.50 43-14 19 and pulse (=wing stroke) rates went to zero at about 4OC when their temperature regressions were extrapolated downward. The same phenomenon can be seen in other insects, as well, including several tettigoniids studied or reported by Dumortier (1963), Shaw (1968), and Whitesell and Walker (1978) and two delphacid homopterans studied by Booij (1982). Temperature data for initial volley frequency and volley repetition rate of lacewing calls are also reasonably consistent with the concept of regression convergence, since the x-intercepts for relevant call parameters range from 2.82OC to 9.40å¡ in Ch. plorabunda and average 8.3 1å¡ in Chr.~ysopa oculata (Henry 1982b). However, the flatter slopes of the temperature- frequency regressions for middle and terminal portions of plora- bunda volleys cause those lines to intersect the x-axis at -1-27OC and -19.09OC, respectively, which does not fit well with Walker's generalization. Moreover, other examples from the literature fail to confirm the phenomenon, even in certain crickets and katydids recently studied by Walker himself (and collaborators): for instance, Anurogr~~llus arboreus and winter races of Neoconocepha/us triops both display rather flat regression lines of temperature versus wing stroke rate that intersect the x-axis well below -5OC (Prestwich and Walker 1981, Whitesell and Walker 1978). Thus, Walker's "four



================================================================================

Figure 4.
Effect on terminal volley frequency (C) of artificially mass-loading the abdomen of four Ch. plorabunda individuals, superimposed on a graph showing the calculated temperature regression lines of terminal volley frequency for 10 other insects of the same species.