Learning to Identify Insect Songs

by Carl Strang

One of the obstacles to a singing insect monitoring program is the large number of various songs that need to be learned for identification. This is not really much different from learning bird songs for breeding bird monitoring, however (except that the total number of species is smaller here). Instead of being daunted by the entire process, it is possible to take the learning process in stages, beginning with the songs that are common and easy to recognize, the ones you have been hearing all along but simply didn’t have the species labels. Here is a list of a dozen suggested species to start with in the first stage: spring field cricket/fall field cricket (their songs are identical), Allard’s and striped ground crickets, snowy tree cricket, common true katydid, black-legged meadow katydid, greater angle-wing, round-tipped conehead, dog day cicada, scissor-grinder cicada, and Linne’s cicada (for more information on these species, try the tags at the head of this post).

Snowy tree cricket, one of the species on the starter list

This list and those that will follow are for northeastern Illinois and northwestern Indiana. There would be substitutions in other parts of the country (I encourage readers elsewhere to make comments here with their own suggestions). Go to reference recordings of these species’ songs, either on-line at the Singing Insects of North America website or through the CD that accompanies the Songs of Insects book. It is not too late this year to hear many of the species on this list on the warmer days, though some are finished or nearly so.

My recommended species list to focus on in the second stage of learning consists of 8 species and groups of species: greenstriped grasshopper, gladiator meadow katydid, Roesel’s katydid ( three species that sing relatively early in the season), and then later, Carolina ground cricket, Say’s trig, sword-bearing conehead, two-spotted/narrow-winged tree crickets (no need to worry yet about separating the two), and the meadow tree cricket group (3-4 species whose songs are essentially identical to the ear and will remain so).

Roesel’s katydid is a species from the second-stage list.

This list of common species either will take you to additional, though still readily available, habitats, or else require a little more of a practiced ear (which practice you got with the first species group). In particular, seek out and spend some time getting familiar with the songs of the Carolina ground cricket and Say’s trig. They need a little more effort to recognize in the field, but once you have them, they will be touchstones for many other species (much as robin songs are for learning bird vocalizations). If you are starting now, you might push the Carolina ground cricket to the first list, as it is one of the few species singing on the cooler days and evenings.

Once you have mastered the second list of species, you are ready for the more subtle distinctions needed to distinguish the songs in the third species list. This includes separating out the song of Linne’s cicada from similar songs by the lyric cicada, and in some areas, swamp and/or northern dusk-singing cicada.

Linne’s cicada

Also, by this point you are ready to distinguish the two-spotted tree cricket song from that of the narrow-winged tree cricket. Also, the broad-winged tree cricket should stand out now from other long-trilling species. In addition, you no doubt have noticed and begun to puzzle out other species that are more idiosyncratic in their distribution or smaller in numbers that you have encountered in your favorite places.

And that brings you to the fourth stage, learning the songs of whatever remaining species may live in the area you wish to monitor. For this you will need a regional guide. In the Chicago region, you can meet this need with the guide I am developing. It is available for free as a .pdf e-mail attachment. Simply request it at my work e-mail address: cstrang@dupageforest.com

As you are learning and listening, pay attention to which songs you can hear clearly, and at what distances, and which are marginal. This will inform the limitations you will need to address or acknowledge in your monitoring.

Block Count Summary

by Carl Strang

Neighborhood block counts are a survey method I have developed in my singing insects study. They consist of walks around the block in my residential neighborhood in Warrenville. I record the time and temperature, and count only the insects whose individual songs I can distinguish from the general background of tree cricket and ground cricket songs. I try to include a mix of times, especially early in the season to include cicadas, but concentrate most of the effort after dark when most species are singing. The overall species count in 2011 was 14 (no new species this year).

One of the streets on the block.

Most of these insects do not occur in high enough numbers to permit statistical comparisons between years. The high counts for each species this year were as follows: 1 for fall field cricket, jumping bush cricket, Say’s trig, and Davis’s tree cricket; 2 for Allard’s ground cricket and dog day cicada; 3 for snowy tree cricket, 4 for narrow-winged tree cricket, 5 for common true katydid, 8 for two-spotted tree cricket, 10 for Linne’s cicada, 14 for Carolina ground cricket, 16 for greater anglewing, and 26 for striped ground cricket. Most of these are typical of the past five years. Apart from statistical comparisons to follow, there seems to have been a decline in fall field crickets over that period (high counts of 5 in 2007, 4 in 2006, and 1 in the recent years, but in 2011 that count was on only one occasion).

Only three species occur consistently in large enough numbers to justify statistical comparisons between years. There never have been statistically significant differences between adjacent years in striped ground cricket counts, and the same was true in this year’s comparison with 2010 (overall medians 14.5 in both years; Mann-Whitney U-test, z = -0.14, P>0.01). Carolina ground crickets likewise were not different (median 4 in 2010, 7 in 2011; z = -2.07, P>0.01). Greater anglewing counts did show a statistically significant increase over last year (median 1 in 2010, 4 in 2011; z = -3.30, P < 0.01), and in general they and the fall field crickets have been the most volatile.

Peak Singing Insect Season Begins

by Carl Strang

By this point in the season we are hearing nearly all the common singing insects in northeast Illinois. First song dates on the whole have been in line with those of recent years, but a few have been relatively late and one was comparatively early.

Snowy tree cricket

My own first observations of singing males typical of residential neighborhoods have included Carolina ground cricket (19 July), common true katydid (also 19 July, a relatively early start for that species), Allard’s ground cricket (22 July), snowy tree cricket (29 July), fall field cricket (30 July, a relatively late start but I had been out of town for a week), two-spotted tree cricket (31 July), Say’s trig (31 July), and greater angle-winged katydid (4 August).

Greater angle-winged katydid

I didn’t get into forests at night until the Roger Raccoon Club campout, so my first date of August 3 for oblong-winged and rattler round-winged katydids has limited meaning. Another forest species that may have started up before my first observation of August 2 was the confused ground cricket.

Black-legged meadow katydid

Our most common large meadow katydid, the black-legged, started up around August 2, a relatively late start for this wetland-edge insect.

Insects at the Campground

by Carl Strang

The overnight campout at Waterfall Glen Forest Preserve that concludes the Roger Raccoon Club provides opportunities to observe various aspects of natural history. Last week the campground’s gravel parking lot, damp from overnight rain, apparently held a good supply of dissolved minerals, as it attracted butterflies representing at least 5 species.

Most remarkable was this giant swallowtail, a species uncommon in DuPage County but encountered regularly in small numbers at Waterfall Glen. I believe their main food plant in northeast Illinois is prickly ash, an infrequent shrub.

After dark an orchestra of singing insects filled the forest with sound. Dominant that night were large numbers of common true katydids, with smaller numbers of scattered two-spotted tree crickets, rattler round-winged katydids, and oblong-winged katydids. I needed a photo of the last species, and was able to find one singing close to the edge of the woods.

This oblong-winged katydid is holding his wings apart in singing position. Rasp and file structures on the brown portions of the two wings are rubbed together to produce the song.

As it happened, some of the kids found another male on their tent in the morning. I gave them a brief lesson about it, and took another set of photos.

Here the wings are in their normal folded position.

Some of those singing insects represented my first observations of their species this year.

Song Peaks

by Carl Strang

My block count data allow me to look at when different kinds of singing insects are most active. I note the time of day for each count, and this year categorized four years’ counts with respect to number of hours before or after sunset.

Two-spotted tree crickets sang the most in the hour after sunset, then continued intermittently through the night.

Sample sizes were 1 count 5-6 hours before sunset, 3 counts 4-5 hours before sunset, 9 counts 3-4 hours before sunset, 13 counts 2-3 hours before sunset, 20 counts 1-2 hours before sunset, 29 counts 0-1 hour before sunset, 61 counts 0-1 hour after sunset, 59 counts 1-2 hours after sunset, 23 counts 2-3 hours after sunset, and 11 counts 3-4 hours after sunset. Some of these sample sizes are small enough to consider the results as tentative (especially for cicadas, which sing before sunset), but on the whole they seem to reflect my more general observations of activity peaks in DuPage County.

Snowy tree crickets peaked 2-3 hours after sunset. Some tree crickets like to nibble on human skin, given the chance.

I found the following peak singing times:  4-5 hours before sunset for Linne’s cicada, 1-2 hours before sunset for the dog day cicada, 0-1 hour after sunset for jumping bush cricket and two-spotted tree cricket, 1-2 hours after sunset for fall field cricket, snowy tree cricket, common true katydid and greater anglewing, 2-3 hours after sunset for striped ground cricket and Carolina ground cricket, and 3-4 hours after sunset for narrow-winged tree cricket.

Narrow-winged tree crickets showed the latest singing peak in my neighborhood, 3-4 hours after sunset.

One particular bias to keep in mind is that some loud singers may drown out others, reducing their counts. For instance, ground crickets are singing at the times when cicadas are active, but I can’t hear them above the cicadas’ loud drones.

A Week of Firsts

by Carl Strang

The past week brought the true beginning of the singing insects season. Though a full work week and other obligations filled my time, I was able to take advantage of what I have learned in past seasons to make some observations that were valuable to me.

For instance, last year I finally nailed down features of the two-spotted tree cricket’s song that removed my earlier uncertainty. The breakthrough came when I found this male, who had chewed a hole in a grape leaf to create a baffle that amplified and/or directed his song.

Finally seeing one of his kind in action, I was able to define the song of that species by the irregular length of its trills, some reaching 7 seconds’ length or more, the spaces between trills being either very brief, or longer but filled with stuttering sounds, and the strained or discordant quality of the sound.

That discovery allowed me to be confident in identifying the first two-spotteds singing in DuPage County, Illinois, this year, on July 12. That is the earliest song date I have for that species, by 5 days.

I also heard the year’s first ground crickets in DuPage. I had been gone a week, and returned to hear striped (photo below) and Allard’s ground crickets on July 12, and Carolina ground crickets on the 13th (I had heard all three near Culver, Indiana, on July 8). Though not the earliest ever, all three species were singing near the earliest first song dates I have noted for them since my study began in 2006.

Mayslake Forest Preserve is one of the DuPage County sites where there are fall field crickets but no spring field crickets. So, when I heard a field cricket song there on July 14 I marked the beginning of that species’ singing season (female shown).

During First Folio Theater’s evening performance of Twelfth Night at Mayslake I heard a few protean shieldbacks singing, my first record of that katydid species on that preserve (female shown).

Finally, I heard my first sword-bearing coneheads of the year while driving to Culver on the evening of July 16.

There are plenty of other species yet to begin, but it feels like the singing insects season has begun in earnest.

Arboreal Tree Cricket Songs

by Carl Strang

This year I finally was able to resolve my confusion about the songs of three arboreal tree crickets. The first breakthrough came from observing singing male two-spotted tree crickets, as I mentioned in an earlier post.

Two-spotted tree cricket (male)

After further study I developed a description of the two-spotted tree cricket song that I feel confident about. Their songs are composed of variable-length trills, 1 second to 7 seconds or longer duration. Spaces between trills can be very short or longer, but when longer usually are filled by a stuttering sound. Often their song seems more discordant or strained than those of the others. Both two-spotted and narrow-winged tree cricket songs are distinctly higher pitched than the song of a common reference, the snowy tree cricket, but not necessarily well separated from one another. When there is a slight difference, the two-spotted has the slightly higher pitch.

Narrow-winged tree cricket

Trill length and the spaces between trills are very consistent in narrow-wings, forming a regular rhythm, unlike the variable two-spotteds. Spaces between always are distinct, longer than the brief pauses most characteristic of two-spotteds, and do not contain stutters. The trills are brief, a couple of seconds long. There was a difference in when the species started, with the first two-spotted singing at the beginning of August, narrow-wings not beginning until the end of the month. Both species continued to the end of September.

On two occasions I heard what could only be a Davis’s tree cricket. The long continuous trill, lower pitched than that of the snowy’s song, occasionally was broken. The cricket was well up in an ash tree. Though I only heard this one, I now suspect that I seldom hear them because their usual location is high in trees, according to the literature, where the higher-pitched chorus of other nocturnal crickets may obscure their songs. The lower pitch and much longer trill matches written descriptions in the literature, if not reference recordings (which give the impression that trills are more interrupted than in fact they are). For reference recordings, go to the Singing Insects of North America website or the Songs of Insects website.

Block Count Results

by Carl Strang

Earlier in the season I described my block counts of singing insects, in which I walk around the block where I live, counting the singing insects I hear. In these counts I heard 13 species in 2009, compared to 15 in 2008. The two missing species were the dog day and scissor-grinder cicadas, both of which I heard on my block at times other than sampling sessions.

Only 3 species have been present in large enough numbers to allow comparisons between years. Most abundant is the striped ground cricket, with median counts of 13, 14.5, and 12 for 2007, 2008 and 2009, respectively*. Statistical tests found no differences among years in the striped ground cricket.

Median counts of Carolina ground crickets were 3, 3, and 7 for the three years. The change from 2008 to 2009 was statistically significant. The third species, the greater angle-winged katydid, produced median counts of 7, 1, and 0. The drop between 2007 and 2008 was statistically significant, but that from 2008 to 2009 was not.

The other species I have noted in my neighborhood, in rough descending order of abundance, are two-spotted tree cricket, snowy tree cricket, fall field cricket, narrow-winged tree cricket, Linne’s cicada, common true katydid, Allard’s ground cricket, and a few species I have heard very infrequently: jumping bush cricket, Say’s trig, Davis’s tree cricket and possibly fork-tailed bush katydid.

*For those interested in the technical details, I included only counts in August-October, the months when these species appeared consistently. Total numbers of counts were 49, 46 and 35 in the three years, respectively. I am very conservative in my use of statistical tests. Here I used the large-sample version of the Mann-Whitney U-test, a nonparametric test which produces (ironically?) a z test statistic. Non-parametric tests are needed for count data, and I seldom see anything like a normal distribution in the aspects of nature I study, anyway. Test statistic values for striped ground cricket counts were z = -0.73 for 2007 vs. 2008, and -1.58 for 2008 vs. 2009, both P > 0.01. Again, being conservative, I use the 1% rather than the 5% level of error. Carolina ground cricket counts showed no change between 2007 and 2008 (z = -0.68, P > 0.01), and an increase from 2008 to 2009 (z = -3.32, P < 0.01). Greater angle-wing counts showed a decrease from 2007 to 2008 (z = -3.33, P < 0.01), but no change from 2008 to 2009 (z = 0.32, P > 0.01).

Gadget 1

by Carl Strang

For the most part in this blog I am trying to model methods of inquiry that don’t rely on technology. Our human senses have their limitations, but we can gather enough information through them to answer a lot of questions about our surrounding wild world. Nevertheless, there are occasions when gadgets can help. Today I will feature one of those I have found useful in my field studies of singing insects: my soprano recorder.

Recorder b

I don’t have perfect pitch, but I have a reasonably good ear. I have found the recorder to be especially helpful as I tackle the problem of the arboreal tree crickets (outlined in my earlier post on one of them, the two-spotted tree cricket ). This season one of my goals has been to sort out the songs of the two-spotted, narrow-winged and Davis’s tree crickets. I was encouraged when I noticed that Elliott and Hershberger, in their recent book on singing insects, indicated that these three species should have distinct pitches in their songs. Highest should be the two-spotted, at 3.5 kHz (kilohertz, a quantitative sound frequency measurement), which translates to a pitch of A, the fourth A above middle C. In the middle should be the narrow-winged tree cricket, at 3 kHz or approximately F-sharp, the fourth F-sharp above middle C, 3 half-tones below the two-spotted. I noticed, incidentally, that the distinctive song* of the snowy tree cricket also is indicated to be at 3 kHz, and so I had hopes that this would provide a rough and ready field standard. The lowest of the arboreal tree crickets, according to Elliott and Hershberger, is the Davis’s tree cricket, at 2.5 kHz. This translates musically to the fourth E above middle C, distinctly lower than any of the others. So, recorder in hand, I set forth.

Snowy tree cricket 5b

Snowy Tree Cricket

Two caveats quickly became clear. First, the pitch of a given species is subject to change with environmental temperature, rising and falling as the temperature rises and falls. Second, I have to keep in mind that my own hearing may not well match the measuring devices used to provide the information in that book. In general all species sounded, to my ear, a good 3 tones lower than Elliott and Hershberger suggested.

I have found that to my ear, both two-spotted and narrow-winged tree crickets have songs distinctly higher pitched than that of the snowy tree cricket. At a given temperature, the two-spotted sings one-half to a full tone higher than does the narrow-winged. However, at a given general temperature, narrow-wings range over half a tone of pitch or more. Whether this is because the microclimate is different where individual crickets are singing, or whether this is something they are controlling, I cannot say. It means, though, that I have to rely as much or more on the temporal pattern of the song to distinguish these two species.

Narrow-winged Tree Cricket

Narrow-winged Tree Cricket

Narrow-wings sing with a steady pattern of trills and spaces, with trills of equal lengths and spaces of equal lengths, and the spaces are significant at a second or so duration. Two-spotteds sing at a little higher pitch on average, have trills of varying lengths including some often lasting well over 5 seconds, usually with at least some pauses that are very brief, as though catching a quick breath.

I recently heard, on my neighborhood block count, what I believe must have been a Davis’s tree cricket. The insect was high up in a tree. Its trills were variable but generally very long, with only occasional odd interruptions. Spaces were short. Significantly, the pitch was down at A-flat, low for the temperature, which was 70F. Based on my recorder tests, at that temperature I would expect snowy tree crickets to be singing at B or C, two-spotteds at the E above that, and narrow-wings at C to E. So, the recorder is a helpful tool, but in distinguishing the songs of these crickets I find that the pattern of their song is more reliable than the pitch.

*You know the song of the snowy tree cricket, even if you live outside its range. In the movies, whenever the director wants to convey a calm nighttime mood, there will be a snowy tree cricket in the sound track. The song is a pulsing tone, varying with the temperature so that if you count the notes in 15 seconds and add 40, you have the temperature in degrees Fahrenheit. Though the narrow-winged tree cricket’s song likewise is a regular pulse, the tones are on the order of 2 seconds’ duration with a 1-2-second space between. Unless the temperature is very cool, the snowy’s song is much faster.

Two-spotted Tree Cricket

by Carl Strang

One of the goals in my singing insects study this year is to sort out the songs of three arboreal tree crickets. In the field I have found that their songs are not as distinct from one another as reference recordings and descriptions seemed to suggest. Two of the three species I have seen, and so confirmed their presence in DuPage County. Today I begin with what I suspect may be the only one singing as early in the season as late July and early August: the two-spotted tree cricket.

Two-spotted tree cricket 1b

This photo shows a female, with the two large spots on her back that give the species its name (males lack them, and are pale). She sits on the arm of one of the 2006 Roger Raccoon Club  kids, who brought her to me for identification. Until two weeks ago, she was the only one I had seen. Certainly the references were correct in saying these are not easy to find. They live in trees, often well above the ground. The male’s song, which you can find here  or here, is a strained, often dissonant sounding trill that is interrupted fairly frequently by brief pauses that often are filled with stuttering sounds. Unfortunately, the same description applies more or less to the songs of Davis’s tree cricket and the narrow-winged tree cricket, though the tone of the last seems more melodic to my ear.

Two-spotteds begin to sing at dusk. On August 6 I was at Timber Ridge Forest Preserve, strolling the Great Western Trail with ears open for insect songs, when scattered tree crickets in this target group began to sing. All had identical songs, but one in particular seemed to be closer to the ground and just off the trail. After a short time I found him.

2-spotted singing b

He was on the underside of a big grape leaf. Here he is close up.

2-spotted singing cropped b

He was using a trick for which some of the tree crickets are known. He had chewed a circular hole in the leaf, and was using it to amplify and possibly direct his song (tree crickets sing by elevating their wings and vibrating them against one another).

2-spotted wings down b

I made a recording, then prepared to collect him for identification. But when I put my flashlight on him again I found this was unnecessary.

2-spotted pair 2b

A female had arrived on the scene, and there was no mistaking her identity. The male kept his wings elevated, and continued to vibrate them occasionally in song. She was palpating her way slowly up his back in search, I believe, of secretions that some of the tree cricket males provide as nuptial food gifts in a prelude to mating.

2-spotted pair 1b

The next evening at dusk I was at Springbrook Prairie Forest Preserve. I heard the same song coming from near the top of a 15-foot-tall bur oak beside the trail. Looking up toward the point from which the sound seemed to be coming, I noticed that one of the leaves had a circular hole in the middle. When I illuminated it with my flashlight, sure enough, there was another male two-spotted tree cricket. So, at least with plants having relatively large leaves, I now know to look for distinctive circular holes that may help me to find these elusive insects.

%d bloggers like this: