The earliest record that I know of peppered
moths is 1776, when an entomologist called Moses Harris
described the black and white insect he could find near London.
He also pointed out that the larvae came in different colours.
Since he noted the variation in the larvae, he was likely to
note variation in the adults if there were any.
I thought I
would just show you some pictures.
about 200 moths that show something you could call industrial
melanism, that is, the presence of dark forms at higher
frequencies in industrialised areas. But the one I am going to
talk about exclusively is the peppered moth (Biston betularia),
which is the most famous and obvious example. I also propose to
describe how ideas on the subject have developed, both among
scientists and the general public.
The speckled form is the sort that Harris would have seen. In
the early nineteenth century there are one or two examples of
black forms in collections but they only became well
established by the middle of the nineteenth century, and the
first black forms which have actually been located are from
Manchester, in the Cheetam Hill area. They look like the next
On a pinned specimen there is a little bit
of the typical black and white pattern, but that part of the
wing would not be seen in life. The rest of the insect is
started to appear in the second half of the nineteenth century.
This is a
picture of a collection taken in Manchester in the early 1970s.
In about 120 moths there are four typicals. The rest are the
melanic form called carbonaria.
There are some
other forms, normally fairly rare, which are intermediate in
colour (called insularia), such as these in the centre.
was published in 1908 by a Manchester man, carbonaria at the
bottom, typical at the top and insularia between.
forms are produced by a single locus. The darker are dominant
to the paler forms. There seem to be three insularia alleles.
They are difficult to distinguish in the field, but sometimes
you can separate collections into the different categories.
So that is
the variation that developed in the 19th century, and excited
people to discuss or write about the variation were amateur
entomologists in places like Manchester. They wrote, in amateur
entomological journals, fairly anecdotal remarks such as “I
went out and made a collection and what do you think I found?
Several species that contained unusual black individuals”.
and comments that it elicited, was not really to do with ideas
about evolution, it was to do with variation. People were
concerned about what species they had. They were interested in
variation because it shed some light on whether or not they
were dealing with a new species.
The new black
forms excited some attention in the amateur entomological
literature. People asked themselves why the change had happened
and, basically, there were three kinds of answer.
1. It may
because the newly industrialised towns are warmer than the
surrounding rural areas, and the dark forms like warm
2. It may be that the larvae are eating something
poisonous (e.g. lead salts) deposited on the trees, that turns
3. Or, it may be that there is a difference in
camouflage and dark colour is a protection against predators.
All of those
ideas were floated in the nineteenth century.
shows what Manchester used to be like at the time we are
talking about. When I came to Manchester the Dunlop factory
building visible though window, which is now apartments, looked
like this. It didn’t have smoke coming out of the top but
we had the chimney and completely black-walled buildings. The
area we are now in was shops. Just behind were narrow streets
with two-up, two-down houses, all of them at that time with
domestic grates burning coal. So there was a tremendous amount
of smoke and general dirt, which was obviously noticeable to
nineteenth century people.
in this topic from an evolutionary point of view came from the
scientific establishment, which was conscious of a problem. The
Darwinian idea was that evolution took place through adaptation
and selection producing imperceptible changes; very slow
progressive modifications in organisms occurred to suit their
environment. The inherited component, which was carried forward
in evolution, was thought at first to be due to some kind of
blending of material that was passed on from one individual to
another. By the 1880s or so, several difficulties were seen
with this explanation.
One is that,
if the inherited material is blended, the distinctions would
disappear, so you had to have some kind of discontinuous
inheritance. This was understood well before Mendel’s
results were discovered by the general scientific public. The
other problem was that species could be seen not blend into
each other; they were discontinuous in characters.
Francis Galton and Frank Weldon set up a committee of the Royal
Society to study, statistically, the continuous variation that
they observed in organisms, with a view to supporting general
evolutionary ideas according to the Darwinian imperceptible
student of Weldon’s, William Bateson, became involved.
Basically what impressed him was that there is a fundamental
discontinuity between species, so we have to look for something
that causes us to jump from one species to another. There were
also other kinds of evidence that the situation is more complex
than the continuous change model. For example, if you look at
pairs of closely related species you often find one of them is
very variable, while the other is not.
Yet they live
in the same kind of environment doing the same kinds of thing.
He quoted various moth species for which that is true. The
implication is that species are organised in some kind of
physiological sense, he said. In other words, to do statistical
measurements alone is to bark up the wrong tree.
To cut the
story short, Bateson became involved with the Royal Society
group. It began as the Committee for the Study of Statistical
Variation in Species; he turned it into the Evolution
Committee. He commissioned people to look for examples of
discontinuous variation, and wrote to amateur entomological
journals saying “we need data, get it together and send
it to me”.
therefore the person who founded the scientific study of
peppered moths: a nice example for him because big changes were
taking place and he thought that this may be important from a
speciation point of view. At about the same time it was
demonstrated that there was a Mendelian basis to the variation,
so there was no need to keep considering polluting lead salts
as a direct environmental effect changing the phenotype. The
forms segregated when you bred them.
evolutionary change and blending inheritance versus
discontinuous change, both genetically and in evolutionary
terms, was the original issue that people argued about in
connection with the peppered moth.
the information together Bateson had it published, and if you
examine it you get a picture which looks like this.
records were dotted around the country. People sometimes said
“where dark forms were once rare they are now very
prevalent”, information as loose as that. But, if you
interpret these records it looks as if there was a spread from
somewhere in north west England, where there was quite a high
frequency of melanics in the 1860s. To north and south of the
country the first records were not until the 1890s. So the
pattern looks like a classical migration outwards of the
By the time
melanics were being noticed in the southern counties the
frequency around Manchester had got to 90% or more, so the
frequency of melanics here continued to rise as the spread took
place in the south. This is interesting because London had been
a very large conurbation and extremely polluted for hundreds of
years. So, it may be simply by chance that a rare mutation got
going in the Manchester area and spread.
equally well have started in London, I suspect. A lot of
seventeenth and eighteenth century literature recorded how
terrible the atmosphere was in London, and it was noted how
much clearer it became once when the Dutch blockaded the east
coast and stopped coal boats coming down from Newcastle.
So London had
the sort of problems we are used to associating with the north,
yet the melanic form was late in increasing in frequency in the
There was a
gap in peppered moth studies in the 1920s, when people lost
interest in the mechanism of evolution as a subject in itself –
they became interested in phylogenies and function instead.
This gap takes us up to about 1950 when the principal person to
enter the field was Bernard Kettlewell.
He had a
grant to study industrial melanism in general and the peppered
moth in particular, to see what the pattern had become and why
it was brought about. At that time he again assembled data on
frequency. They were originally published as pie diagrams,
which show the limits of the records better than a contour map,
but contours provide a neater pattern to look at.
Kettlewell data come from three surveys made between 1952 and
1972 - about 20 years. It didn’t look as if any change
had occurred over those 20 years. The environment did not
change much either, so the whole data set looks like a static
pattern with high melanic frequency in industrial regions and a
very steep cline going into rural N. Wales.
The change to
the south is less steep. The whole of the Liverpool and
Manchester area was over 90% melanic [carbonaria] phenotype
frequency. There was an area of high insularia in south Wales
[lower left], but this map just shows carbonaria frequencies.
got together a large collection of records for the middle of
the twentieth century; a much better set, but similar in
pattern to that collected at the end of the nineteenth century.
It suggested that, while in the nineteenth century it had been
changing, by the beginning of the twentieth century the pattern
had become a fixed.
decided that one of the important things to look at, in trying
to explain why the pattern occurred, was camouflage and
selective predation. There are various reasons for choosing
this approach that we won’t go into at the moment, but
one important point was that camouflage and selective predation
were susceptible to experimental work.
started a series of experiments. First he put insectivorous
birds in cages and fed them moths of different kinds. While
doing so, he judged whether he could see the moths or not, and
the ones that he couldn’t see well, the birds couldn’t
see very well either. This gave him confidence in the story.
experiments showed that birds actually will eat moths,
something which had been argued about previously. That
suggested to him that it was worthwhile carrying out a field
experiment. This he famously did, with one set of trials in
Birmingham and one in Dorset. The atmosphere of Birmingham was
extremely polluted at the time, that of Dorset unpolluted. Both
experiments were in woodland. In Birmingham most wild peppered
moths were melanic and in Dorset he found no carbonaria at all.
mixtures of different forms of live moths and recaptured them.
When you do that you know the frequencies released and the
frequencies that you recapture. It is then possible to compare
the one with the other to get an indication of the relative
survivability of the two types.
conditions involved in the tests were extreme – this is a
black and white story. A picture of moths in North Wales on a
characteristic piece of tree trunk shows clearly that the
carbonaria form is more obvious than the typical.
contrasting picture of moths on an ash tree in my back garden
in Rusholme in the early 1970s shows that there typicals were
much more conspicuous.
surprisingly, given the differences involved, carbonaria was
twice as likely as typical to survive per day in Birmingham and
half as likely to survive per day in Dorset. Niko Tinbergen
took a photograph, in Birmingham, of a cock redstart with a
typical in its beak that was graphic evidence of predation. So
‘Hey Presto' we have the answer to why the frequency
increased! That very much became the accepted, the ‘text
there was a context within which this experimental work was
carried out. Other people, such as Tinbergen, were doing
similar experiments. Tinbergen was interested in studying
ethology, the behaviour of animals in the field. He introduced
an experimental approach that showed you could learn a
tremendous amount about behaviour patterns in species such as
gulls and sticklebacks
to his direct interests, he was also opening up a type of
experimental procedure that showed that simple experiments
could produce quite profound answers. It allowed judgements to
be made about evolutionary processes. Before that time people
thought of evolutionary studies as measured by phylogenetics or
geological change. The evidence involved slow processes from
which you could make inferences, but with which you could not
empowered by the new type of experiment, to find things out in
the field in quite simple ways. It was a ‘heady’
feeling at the time. Kettlewell belonged to this group of
experimentalists and worked directly with Tinbergen. That is
one of the reasons why the peppered moth story became such an
important example. It was topical in the way it was handled.
shows survey work carried out by Clarke and Sheppard and Bishop
in the Liverpool, N. Wales area, and by myself and others in
Manchester, in the late 1960s and early 1970s. The trouble with
Manchester as a study site at that time was that nothing much
happened. The frequencies were all very much the same, whereas
in N. Wales we have this very steep cline. It is useful because
it allows us to do experimental manipulations. You can imagine
that the selection is likely to vary considerably across short
distances there, and quite a number of experiments, along the
lines of Kettlewell’s, were carried out.
A couple of
general surveys have been done since. The first was undertaken
by Open University students. The students were asked to collect
moths in traps and send examples to be identified. Once we had
checked them and thrown away those that weren’t peppered
moths it produced this picture in 1982-1984. The central core
of high frequency had contracted and the area to south was
broken up into more irregular patches, but the original
Kettlewell pattern can still be seen.
in the mid 1990s, there was another survey using data obtained
by the Rothamsted Insect Survey. They had been collecting moths
all over the country and by the mid 1990s we had a pattern like
this, where the 90% frequency patch has disappeared. The
highest frequencies are 50% - 60% and the whole pattern is
flattening. There is still a cline in N. Wales, with higher
frequencies to the N.E, but the whole plateau, which once
looked permanent, is now disappearing rapidly. It looks as if
we are dealing with a process that is a blip in the life of the
peppered moth. Melanics rose in frequency during the industrial
period based on coal burning, then went down.
I am running
ahead slightly so far as the development of ideas about the
subject is concerned. The declines in frequency in the peppered
moth, which took place in concert with manifest changes in the
environment, occurred at a time when there was a lot of
interest in the environment, particularly with problems of
(some of them in the 1970s with long hair, flares and flowered
shirts), were deeply pessimistic about the future of the
environment. But here was a storey to be optimistic about in
the field of environmental biology. It showed you could improve
the environment and that there could be a response as far as a
species was concerned.
the change, this is a picture of Manchester University tower in
the early 1970s. When the smoke control legislation was
established and all of the small houses around here were pulled
down, the university decided to clean its buildings.
This is a
test area on the stone work. Subsequently the authorities went
ahead and to produce a beautiful, clean, optimistic looking,
environmentally friendly, seat of learning.
moth story therefore showed that environmental improvements
could be followed through with responses by natural organisms.
This was a matter of public interest, which generated quite a
lot of excitement at the time.
This is a
summary diagram of change in frequency over time. There are two
ways in which we can try to measure the forces that must
operate to produce the changes that have taken place. One is
directly, as Kettlewell did, the other is to look at points in
a sequence. If you have a series of values then you can
estimate the amount of selection required to cause the change.
This is a summary for the Manchester situation.
in blue at the left are from general impressions people had:
“when I was collecting at this date I never saw any
melanics, later they were nearly all black”. That is the
sort of evidence we have from the nineteenth century. The
points in red are better records that we have collected over
the twentieth century, including Mike Dockery’s [in the
audience]. His is a small sample with a big standard error.
Nonetheless, it fits in the right place as far as the general
trend is concerned.
shows a longitudinal section through a hundred and fifty years
of change. It really does show a dramatic change.
One of the
most intriguing developments recently is the way the story has
been picked up by the public and press.
been nice improvements to the environment accompanied by
changes in the peppered moth, then we get this type of thing
from the Daily Telegraph. “Scientists pick holes in
Darwin Moth Theory”. Although Darwin had nothing to do
with the moths his picture was attached. The article says
“scientists admit they do not know the real explanation
for the fate of Biston betularia”. Well, they never said
they knew all of it. They just thought they had made some
progress. The article describes the work as worthless, and
other hostile and violent words have been used in other
comments. To find out more you are directed to a creation web
present a very different picture from the one that I have put
forward. I got this from one such web site – for the
Institute for Creation Research. It says about the peppered
moth, “what a wonderful time to be a creationist when
even the supposed best proof of evolution in action is so
flimsy it cannot stand the test of truth”. What are they
been arguments between different biologists about Kettlewell’s
experiments. For example they have been criticised because he
put the experimental moths low down on tree trunks. Well, he
put them where he could see them and photograph them. When you
study them in detail you realise that a naturally settling moth
goes through quite a complex behavioural pattern, as you would
It will land
on a tree, tend to walk up it, come to branches and either
settle under a branch or walk along the branch. It is not then
sitting where the photographs were taken, but in a more
protected position. This is the information that the press
interpreted as indicating that the whole subject has collapsed.
press has also described photographs of moths on tree trunks as
fakes, because they were taken in more exposed places than
those where the moths often rest.
make one ask why the idea of demolishing this topic should have
such an appeal today. You could argue that demonstration is
simple, even almost boring, yet some people feel a need to
I put in this
slide because, although the critics attack Kettlewell’s
experiments, he was not the only person to carry them out. The
slide shows published data from the literature. The fitness of
the melanic carbonaria (measured as w, the frequency remaining
divided by frequency presented) is plotted against estimated
frequency of typicals where the experiment was carried out.
square in the bottom right is Kettlewell’s Dorset result,
his figures for the industrialised area (Birmingham) are open
squares at top left. Later studies were done at different
times, and frequencies in the areas will have changed, so the
pattern is relatively complicated. But fitnesses (w) are
independent of the frequency estimates. The diagram shows that,
overall, the estimated fitness of carbonaria was highest in
places where melanic frequency was high, and vice versa.
who write the critical literature say that Kettlewell's results
are worthless because they were done in the wrong place. But,
there is a whole series of people who have done other
experiments which, to me, support the original argument.
of Kettlewell’s results claims they are worthless because
they were done on the wrong part of the trees. But there is a
whole series of other experiments which support the original
argument. Some were by Howlett and Majerus (open circles), two
researchers who studied natural settling of the moths and
queried the original experiments. They carried out experiments
in the “wrong” and the “right” place –
which produced the same results. Two other tests by Lees and
Creed, (crosses on the graph) show that melanics are better
protected on dark, wet tree trunks than pale, dry ones. Such
results support the Kettlewell argument.
to left of the graph represent tests carried out by Clarke and
Sheppard in Liverpool. They used dead moths, which allowed them
to classify the sites where individuals were attached. Dark and
light moths were placed in dark and light patches in all
types were on pale backgrounds, carbonaria was lost more than
typical. For the others carbonaria had the advantage. There is
little doubt looking at the graph that where the melanic
frequencies were high there is an advantage overall to
carbonaria. The reverse is true where melanic frequency is low.
Do you believe it? This evidence was simply rejected by Judith
Hooper, for example, who has written a book effectively
differ in their accuracy and information content. Since it is
important to establish the probability of the trend I decided
to carry out a significance test that was as robust as
possible. Because figures on both coordinates are subject to
error I decided that the safest approach was to convert fitness
values into normal deviates (by dividing the deviation of each
estimate from equal fitness by its standard error). You get the
set of points in the next graph.
are significant if they are more than two standard deviations
from the 0 value, where there is no selection. At top left
selection is in favour of carbonaria, at bottom right selection
is against it. The mean deviates for the left hand third and
the right hand third of points are about +3 and -3, giving a
Chi squared value of about 9 for each group. To estimate the
significance of the trend we should add them together, giving a
Chi squared of about 18 with 2 degrees of freedom. That is a
highly significant result. It is clear that the trend cannot
simply be rejected, as the ‘anti’ lobby has done.
Part of the
criticism may be due to a contemporary reaction to reductionist
science. To try to understand the problem we could describe the
Kettlewell programme as in some respects naïve. It says.
You can measure selection.
• What you measure
• What you measure is going to be important in
• Change in gene frequency, as seen here, is
fundamental to evolution.
• Once it is demonstrated
that changes in gene frequency can occur rapidly then all the
other evolutionary processes follow.
But this is a
limited view of evolution. To understand the full range of
phenomena we observe and which amaze us we need much more
complex models. It may be that the approach crudely described
here as a list of bullet points is perceived as arrogant and
limited in its explanatory power. Even so, the validity of the
evidence is unaltered by changing attitudes to scientific
I want to
consider now some research that has been carried out since the
classical peppered moth story was established. The two later
maps of morph frequency that I have shown you are part of the
programme. We can take a look a few other examples.
Clarke and his wife collected the most complete longitudinal
set of records of frequency available. They worked on the
Wirral peninsula, near the edge of the N. Wales cline, sampling
every year from 1959 (black points). A less good data set from
Manchester shows similar changes to theirs (red points). There
are one or two other results like this.
ones come from Kent (SE England, near London), the blue crosses
show figures from Nottingham (E. Midlands). They all present
similar pictures of sharp decline in melanic frequency. When
you look in more detail, the curves with lower starting points
appear to be shallower than those with higher starting points.
This is not easy to see on the curves, but when we examine all
available figures comparing samples from Kettlewell’s
time to recent ones there is a convincing trend.
I worked out
the selective values (w). There is a problem looking at morph
frequencies because the selective value will produce a
different change in morph frequency depending on the start
point. Therefore, the selective values have to be standardised.
The ones which started out with the highest frequency of
melanics [left side] are the ones with the current greatest
disadvantage to carbonaria.
that selection is not uniform over the whole of the country. It
is most intense in the industrial regions where melanics
reached their highest levels and where the environmental
clean-up effort has been greatest.
can look at some evidence relating to insularia. These are the
collecting sites used by the Rothamsted Insect Survey.
I can use the
map to point out that in S. Wales and Gloucestershire there has
always been quite a high insularia frequency. We don’t
know why that is. S. Wales is industrialised and
Gloucestershire relatively non-industrialised. The frequency of
insularia is sometimes higher than that of carbonaria in these
areas, compared with the 1% - 5% found in the northern
‘carbonaria’ regions. Why is that?
reason for the difference is that special conditions in the
south west favour insularia rather than carbonaria, but there
is no evidence what these conditions might be. Another possible
reason relates to the dominance relations of the alleles
concerned. If insularia is intermediate in fitness between
typicals and carbonaria, then at the beginning of
industrialization both carbonaria and insularia would have an
advantage over typical, and increase in frequency. But as the
change proceeds carbonaria becomes the most common morph,
against which insularia is disadvantageous and therefore
decreases. So, the expectation is that curves of changing
insularia frequency have a hump in them. Different starting
frequencies can lead to different patterns of change in
different places. Here is a little bit of theoretical
On this graph
I have plotted the insularia frequency (y) on the carbona
frequency (x). The typical frequency is what is left. The
insularia form is assumed to have a fitness intermediate
between the other two.
left there is 100% typical, and on the diagonal there is 0%
typical. If you start with the same insularia frequency but
different carbonaria:typical ratios the trajectory that
insularia follows varies widely, as shown by the dotted green
lines. It may rise to a high frequency before dropping off, or
it may hardly change at all. That is simple due to the dynamics
of a 3-allele system.
The amount of
selection determines the rate at which the curves are followed,
but not their position, which is determined by the frequency of
the other two alleles against which insularia is compared.
So, it is
possible that the reason why there is high insularia in S.
Wales is simply that insularia became established before
carbonaria did. If conditions remained constant insularia would
eventually decline, to be replaced by carbonaria.
looking at a non-stable situation, however, and all melanics
have subsequently declined.
This is just
an example of the type of inference that can be extracted from
the available data – so long as you do not believe that
the whole story is a hoax!
Would you like to answer some questions? Well, you will answer
Cook: I want an answer to the question why people hate so much
these days. There have been two books written recently, one of
which is entirely about how scientists have tried to perpetuate
a fraud, the other with similar allegations. Why?
is a more comprehensive coverage of this topic.
2003 The rise and fall of the carbonaria form of the peppered
moth. Quarterly Review of Biology 78, 399-417.