scientific correspondence
Step
22
32
1,870
9.860 (26)
0.03049 (53)
26,400
0.18836 (33)
1,317 (56)
62
71
11.785 (22)
0.05191 (40)
940
0.18836 (31)
15,700 (900)
102
24
12.476 (23)
0.05890 (38)
290
0.18790 (57)
24,800 (1,100)
152
5
12.446 (49)
0.05860 (66)
56
0.18784 (51)
23,800 (1,800)
402
17
12.438 (26)
0.05839 (41)
240
0.18785 (58)
22,500 (1,100)
1002
8
12.270 (41)
0.05580 (48)
20
0.18910 (33)
18,900 (700)
4002
2
12.530 (170)
0.06020 (220)
30
0.18831 (57)
21,600 (1,200)
9.800 (77)
0.02900 (20)
0.18800 (30)
295.5 (0.5)
13.800 (100)
0.03280 (50)
0.1724–0.1786
Atmospheric
Solar
Ne
20
Ne/22Ne
21
Ne/22Ne
36
38
Ar/36Ar
Ar
40
Ar/36Ar
~0
Step values refer to the cumulative number of strokes. Numbers in parentheses indicate 1s error in the last digits;
errors in abundances are 5% (1s). Values for atmospheric and solar composition from ref. 1 are for comparison.
NATURE | VOL 399 | 17 JUNE 1999 | www.nature.com
© 1999 Macmillan Magazines Ltd
60,000
Solar 38Ar/36Ar (ref. 1)
0.175
Mantle 40Ar/36Ar (ref. 5)
40,000
40
30,000
38
40
Ar/36Ar
50,000
Ar/36Ar
Ar/36Ar
0.180
Ar/36Ar
Table 1 Neon and argon abundances (210–12 cm3 per gram at standard temperature and pressure)
and isotopic ratios in MORB 2P43
a
38
20,000
0.185
10,000
Air
Air 10
11
12
20
Ne/22Ne
13
b
0.190
Ar/36Ar
Pepin1 suggests, mainly on the basis of three
studies2–4, that solar argon, krypton and
xenon might be present in the Earth’s mantle. But I believe that the analytical evidence
is still too weak to claim such a breakthrough in rare-gas geochemistry. Moreover, other studies5–8, and the new data I
report here, indicate that solar argon, krypton and xenon, if present at all, are only a
minor constituent of mantle rare gases.
Neon isotopes in samples from the
Earth’s mantle clearly indicate a solar-like
composition (see, for example, ref. 1) once
the effect of atmospheric contamination is
removed. Models that explain this feature
also involve the presence of solar-like argon
in the mantle, which should yield lowerthan-air 38Ar/36Ar ratios.
Burnard et al. 2 took laser measurements
of CO2, 4He, 36Ar and 40Ar in a volatile-rich
mid-ocean ridge basalt (MORB) glass
(2PD43) and inferred a solar-like elemental
abundance pattern for all rare-gas species in
the mantle. However, Moreira et al.5
analysed all rare-gas isotopes in the same
sample and showed that this abundance
pattern is close to the atmospheric (planetary) curve, even when corrected for air
contamination by extrapolation to ‘pure’
solar neon (20Ne/22Ne413.8). Valbracht et
al. 3 found low 38Ar/36Ar ratios correlating
with non-atmospheric 20Ne/22Ne and
40
Ar/36Ar ratios in submarine basalts from
the Loihi hotspot, and claimed the first but
“still preliminary” evidence for solar argon
in the mantle. However, if Pepin’s range for
solar argon (0.1724–0.1786)1 is used, the
uncontaminated 40Ar/36Ar ratios deduced
from 20Ne/22Ne versus 40Ar/36Ar and
38
Ar/36Ar versus 40Ar/36Ar systematics agree
less well, which may weaken the evidence
in ref. 3.
The significance of the low 38Ar/36Ar
ratios found in ‘plume on the ridge’ samples 4 is even less clear, because they do not
correlate with elevated 40Ar/36Ar: they are
also observed for atmospheric 40Ar/36Ar
the more primitive source of the hotspots.
Using the data and estimate for the uncontaminated 40Ar/36Ar of the source from ref. 7, a
fairly similar lower limit for 38Ar/36Ar
(0.1868) is obtained.
A possible explanation for these observations is that most 36Ar and 38Ar present
in the mantle may have been re-injected
from the atmospheric reservoir into the
mantle, for example by subduction.
Although subduction-related volcanism
may be an effective barrier for rare gases10,
even a small amount of atmospheric argon
could dominate (pollute) the mantle budgets. If this model is true, then I can derive
from my lower limit for mantle 38Ar/36Ar
(calculated above) that at least 83% of these
isotopes in the MORB source result from
argon recycling by subduction. The first
38
Is there solar argon in
the Earth’s mantle?
ratios. I therefore agree with the authors of
ref. 4 that their data are “compatible with
the air ratio” and that any deviation
observed is hardly significant (S. Niedermann, personal communication).
I am sceptical about this evidence when
I look at the results of other more sensitive
studies6,7 in which no significant deviation
from the atmospheric 38Ar/36Ar ratio was
observed that was accompanied by elevated
20
Ne/22Ne or 40Ar/36Ar ratios, even though
the technique used could have detected
such variations if the effect in the Loihi data
were real. Solar xenon also appears to be
negligible in the MORB source8.
I have studied popping rock (2PD43) to
look for solar argon in the MORB source
(Table 1), taking care to avoid impurities,
interference and mass-fractionation effects,
which could have biased the data. All of our
fractions except one reveal clearly nonatmospheric neon isotope ratios close to
20
Ne/22Ne412.5 and 21Ne/22Ne = 0.06, and
40
Ar/36Ar ratios correlate closely with them,
as described5. But no correlation for the
38
Ar/36Ar ratios ( Fig. 1a) could be found.
A more rigorous argument can be
derived from an argon three-isotope plot
(Fig. 1b), in which two-component mixing
gives a straight mixing line instead of
hyperbolic curves in four isotope plots. A
York fit9 reveals no slope significantly different from zero (a40.1882250.00027,
b4(0.75251.726)21018), which may
indicate that there is no solar argon in the
MORB source. Even taking an extreme
choice of the confidence band (a12s,
b12s), the 38Ar/36Ar ratios for reasonable
estimates of MORB 40Ar/36Ar remain close
to the atmospheric value. By using a
40
Ar/36Ar value of 44,000 for uncontaminated MORB5, the corresponding 38Ar/36Ar is
estimated to be greater than 0.1858 (likelihood¤97.5%). This value remains clearly
distinct from solar 38Ar/36Ar, even for higher
MORB 40Ar/36Ar. However, the best estimate derived from our data is atmospheric,
no matter which mantle 40Ar/36Ar value
is assumed.
I therefore think that solar argon, if present at all, represents only a negligible constituent of the upper-mantle non-radiogenic
argon budget. This also seems to be true for
Air
0.185
0.180
0.175
Solar
10,000 20,000 30,000 40,000
40
Ar/36Ar
Figure 1 Isotope correlations. a, The 38Ar/36Ar versus
Ne/22Ne diagram (green circles, right axis) shows
no correlation, in contrast with the close correlation
between 40Ar/36Ar and 20Ne/22Ne (blue triangles, left
axis). Scaling was adjusted to reach an approximate
overlap for the endmember compositions, atmospheric (40Ar/36Ar4295.5 and 38Ar/36Ar40.188) on the
left and MORB source (40Ar/36Ar444,000 (ref. 5) and
suggested solar 38Ar/36Ar (ref. 1)) on the right. In such
a four-isotope plot, two-component mixing gives a
hyperbola. Such a hyperbola approaches a linear
trend where the ratio of the normalization isotopes
(here 22Ne/36Ar) is equal in the two mixing reservoirs.
This hyperbola is from ref. 5, in which
(22Ne/36Ar)uppermantle/(22Ne/36Ar)air41.650.1. Its low curvature is in agreement with my 40Ar/36Ar data, but it fails
to connect the measured 38Ar/36Ar ratios with the
solar composition. b, The 38Ar/36Ar versus 40Ar/36Ar
diagram shows no correlation, indicating that atmospheric and mantle reservoirs have the same 38Ar/36Ar
ratio. In such a three-isotope plot, two-component
mixing is represented by a straight line, which can
be fitted more reliably than a hyperbola.
20
649
scientific correspondence
direct evidence for such a ‘polluting’ mechanism might come from a correlation
between maximum 40Ar/36Ar with radiogenic lead isotope ratios in MORBs11.
a
b
Joachim Kunz
Institut de Physique du Globe de Paris,
4 Place Jussieu, 75252 Paris Cedex 05, France
e-mail: [email protected]
Singing and hearing in a
Tertiary bushcricket
Communication organs are poorly represented in the fossil record, so their evolution is usually reconstructed by comparison
of extant species using a phylogenetic
approach. We have analysed some extremely
well preserved stridulatory and hearing
organs of the oldest known bushcrickets
from the lowermost Tertiary sediments of
Denmark (55 million years old). These fossils indicate that males sang with a broadband frequency spectrum, and it is likely
that both sexes could hear ultrasound. The
fossil wings have lower asymmetry than
extant species, indicating that bushcrickets
may have evolved from a bilaterally symmetrical ancestor.
Only a few, poorly preserved male
forewings from fossil bushcrickets have
previously been described1. Our specimens
are from Pseudotettigonia amoena, which
was a large bushcricket with a wing span of
more than 120 mm that lived in subtropical
Scandinavia during the Palaeogene2. They
include 11 left and right male forewings and
three forelegs with the ear.
The evolution of courtship behaviour,
with males producing complex songs, has
led to a strong left–right asymmetry of the
forewings in extant male bushcrickets3.
During singing, the left wing is always held
on top of the right, and the ‘stridulatory
file’, which is a dense row of cuticular teeth
on the ventral surface of the left wing, is
scratched over a raised vein on the right
forewing. A thin membrane called the ‘mirror’, which is important for sound radiation
and is surrounded by a frame of strong
650
af
m
c
d
25
Frequency (kHz)
1. Pepin, R. O. Nature 394, 664–667 (1998).
2. Burnard, P., Graham, D. & Turner, G. Science 276, 568–571
(1997).
3. Valbracht, P. J., Staudacher, T., Malahoff, A. & Allègre, C. J.
Earth Planet. Sci. Lett. 150, 399–411 (1997).
4. Niedermann, S., Bach, W. & Erzinger, J. Geochim. Cosmochim.
Acta 61, 2697–2715 (1997).
5. Moreira, M., Kunz, J. & Allègre, C. J. Science 279, 1178–1181
(1998).
6. Poreda, R. & Farley, K. Earth Planet. Sci. Lett. 113, 129–144
(1992).
7. Marty, B. et al. Earth Planet. Sci. Lett. 166, 179–192 (1998).
8. Kunz, J., Staudacher, T. & Allègre, C. J. Science 280, 877–880
(1998).
9. York, D. Earth Planet. Sci. Lett. 5, 320–324 (1969).
10. Staudacher, T. & Allègre, C. J. Earth Planet. Sci. Lett. 89,
173–183 (1988).
11. Sarda, P., Moreira, M. & Staudacher, T. Science 283, 666–668
(1999).
20
15
fossil species
10
5
0
0
50
100
150
Dorsal field area (mm2)
Figure 1 Comparison of the Lower Tertiary Pseudotettigonia amoena with a Recent Tettigoniidae. a, Dorsal
view of Pseudotettigonia left forewing base (specimen HM 14M-3022) showing the stridulatory apparatus.
b, Dorsal view of Recent Decticus verrucivorus left forewing base, showing the mirror (m) and the active file
(af). Scale bars, 1 mm. c, Relation between dorsal field area and dominant frequency peak (y493.886x10.6151,
R 240.79). Data include species from extant Tettigoniidae, Phaneropteridae and Ephippigeridae4–6. d, Tibial
tympanum of fossil Pseudotettigonia (right, specimen HM 14M-C3272) compared with an anterior tympanum
of Recent Phaneroptera falcata (left). Scale bars, 0.5 mm.
veins, is expressed more strongly on the
right wing than on the left, and a complex
of stout spines occurs exclusively on the
dorsal surface of the right wing.
The fossil forewings of males show all
these structures of their extant descendants
(Fig. 1a,b) and are also asymmetrical. The
stridulatory file is much more pronounced
on the left wing, indicating that only the left
file was used for sound production. The left
mirror of Pseudotettigonia is larger than the
right one, although the difference in size is
less pronounced than in extant species. The
spine complex occurs on the dorsal surface
of both wings, in contrast to all living male
bushcrickets. We interpret this condition as
a primitive state in the development of wing
asymmetry, and it is possible that Pseudotettigonia was able to fold the wings back
in both left-over-right and right-over-left
positions.
Comparative morphometric analysis of
the dorsal fields of the fossil wings (the part
of the wings that covers the body dorsally
when the wings are closed) with data from
Recent species4–6 indicates that Pseudotettigonia produced a broadband frequency
song with a dominant frequency peak at
about 7 kHz (Fig. 1c) and an ultrasonic
range, which was probably less pronounced
than in extant species, as Pseudotettigonia
had a comparatively small mirror area. The
size of the mirror has increased during the
evolution of the stridulatory structures,
probably to increase the efficiency of ultrasound radiation4,7. In female bushcrickets,
© 1999 Macmillan Magazines Ltd
most of which are silent, the mirror is
absent.
The ear in the foreleg of Pseudotettigonia
resembles the structure of modern
Phaneropterinae with open tympana8 (Fig.
1d). Although the internal part of the hearing system is not preserved, we conclude
from the modern arrangement of the different areas in the ear9 that the hearing range
was adapted to its own song frequencies, as
it is in extant species. As the fossil bushcrickets could therefore presumably hear at
least low ultrasound, they should also have
been able to hear the echolocation calls of
bats, which first occur in the fossil record at
the same geological age10.
Jes Rust, Andreas Stumpner,
Jochen Gottwald
Insitut für Zoologie und Anthropologie der
Universität Göttingen, Berliner Strasse 28,
37073 Göttingen, Germany
e-mail: [email protected]
1. Sharov, A. G. Trudy Paleontol. Inst. Akad. Nauk SSR 118, 1–217
(1968). (English translation: Phylogeny of the Orthopteroidea;
Israel Progr. Sci. Transl., Jerusalem, 1971).
2. Larsson, S. G. Bull. Geol. Soc. Den. 24, 193–209 (1975).
3. Schumacher, R. Zool. Jb. Physiol. 82, 45–92 (1978).
4. Keuper, A., Weidemann, S., Kalmring, K. & Kaminski, D.
Bioacoustics 1, 171–186 (1988).
5. Kalmring, K., Keuper, A. & Kaiser, W. in The Tettigoniidae.
Biology, Systematics and Evolution (eds Bailey, W. J. & Rentz,
D. C. F.) 191–216 (Springer, Berlin, 1990).
6. Heller, K.-G. Bioakustik der Europäischen Laubheuschrecken
(Margraf, Weikersheim, 1988).
7. Bailey, W. J. J. Exp. Biol. 52, 495–505 (1970).
8. Rentz, D. C. F. Aust. J. Zool. 27, 991–1013 (1979).
9. Bangert, M. et al. Hear. Res. 115, 27–38 (1998).
10. Stucky, R. & McKenna, M. C. in The Fossil Record Vol. 2 (ed.
Benton, M. J.) 739–771 (Chapman & Hall, London, 1993).
NATURE | VOL 399 | 17 JUNE 1999 | www.nature.com