Tuesday 23 March 2021

Frogs: Climate and Spawning. R Maxwell Savage: The Forgotten Doyen of British Ecological Herpetology Part I

Frogs in Ayrshire were spawning last week, a few days before World Frog Day 2021 but unchanged over the past 40 years. Continuing the theme of what external events trigger amphibians to breed, I wondered what had happened further after the extensive research of R. Maxwell Savage on the Common Frog, Rana temporaria in Britain from the 1920s to the 1950s. Savage was something of a mystery man as far as I was concerned. Not only to me I discovered because I was delighted to find that Trevor Beebee had found himself in a similar position. Trevor appealed for information and the result was contact with the Savage family and a short biography in a 2010 issue of Herpetological Journal. I will return to Savage’s life and achievements in a later post but I should say now that I have him down, using my grandsons’ terms, as a super-hero in British herpetology.

I remember reading Savage’s book, The Ecology and Life History of the Common Frog (Rana temporaria temporaria) rather quickly around 1963-64, three years after it was published; quickly because it had to be returned to the library. However, I do remember being aware that one of its main conclusions, that earlier spawning is associated with higher rainfall in the month or so before the actual event, I knew of before then. But how did I know and how did the person who told me know when I could remember being told the story before the publication of Savage’s book in 1961?

Some conversations and their locations stick in the mind while other events disappear from the memory banks. In summer 1959, several of us were in the Junior Biology Lab of the former and now completely demolished Henry Mellish Grammar School and with nothing much to do after taking ‘O’ levels were surveying the biodiversity (in modern terminology) of the school pond. We were talking about tadpoles and frogs when the senior biology master James John Key, who died in 1976 at the young age of 59, told us the story he had picked up about rainfall and spawning. Only when I was doing online searches on Savage did I remember how Jim Key had found out: Ronald Maxwell Savage was on the BBC Home Service radio programme The Naturalist on 8 March 1959. In the programme, edited and introduced by Maxwell Knight, ‘R. Maxwell Savage shows how the spring emergence and spawning of the common frog are stimulated by changes in the weather’ (Radio Times 5 March 1959).

Jim Key listened to that programme and to others on science, not only to keep himself informed of developments but as a source of ammunition for his arguments in the staff room with Stanley Revill, the senior history master, antiquary and archaeologist, who would ask provocative questions in order to stimulate conversation on weightier matters than who had last used the ink eradicator. For example, both listened to P.B. Medawar’s Reith Lectures in 1959, The Future of Man. By this time I was in the Lower VIth and Jim would appear, looking for arguments against (preferably) or for Stan’s latest pronouncement. ‘Natural selection has ended’ was one that kept us going right to the staff room door as Jim re-entered the fray.

After that digression, I want to consider what Savage had done that led him to the conclusion on the importance of rainfall. In Britain spawning date varies with geographical location and from year to year. Savage tried to determine the cause of these variations. It was only a part of his research on the Common Frog but it is an interesting one—and an important one given the effect of climate and weather on natural events. His paper was published in 1935 but he expanded on that exercise in his 1961 book.

Savage’s study of the spawning date of Common Frogs was an exercise in phenology—the study of the times of recurrence of natural phenomena and in particular the influence of climate on plants and animals. Britain has long been obsessed by such events as hearing the first cuckoo in spring or the first flowering of a plant. An enormous ‘citizen science’ project ran between 1891 and 1948 organised by the Royal Meteorological Society with up to 600 recorders submitting returns in some years. Each year a Phenological Report appeared. In the earlier years, the date of frogs spawning was reported but not included in the annual report. Savage searched the individual returns so that he had as complete a set of data as possible.

Savage produced this map from his analysis of phenological reports

Savage then extracted seven meteorological readings from the Monthly Weather Reports compiled by the Meteorological Office. This was a massive task and he received special permission to borrow each volume of reports to work on at home during the evenings and weekends. In these days before computers, Savage entered the data by clipping punched cards (2,734 in all) and then sorted them for each variable by passing a needle through the relevant hole and allowing those clipped to fall out of the stack. In that way, the standard method in statistical operations of the time, he built up tables for analysis and ‘joint functional regression diagrams’.

The diagrams he produced need a little explanation. He was able to plot on a graph, with average monthly rainfall and average monthly temperature as the axes, points at which spawning occurred on certain dates (expressed as day of the year, eg 1 January = 1 etc). He then joined equal dates of spawning with a line—an isophene—analogous to an isobar on a weather map. Each diagram then had a family of isophenes ranging from early spawning (day 20, i.e. 20 January) to very late spawning (day 90, 31 March).

To digress for what he would have obtained had the date of spawning been simply associated with temperature or rainfall, I show theoretical diagrams below. In each case the isophenes would be parallel to one axis or the other. But Savage did not obtain isophenes parallel to an axis. The isophenes were curved showing a complicated relationship between temperature, rainfall and spawning date.

Savage's Diagrams showing isophenes of day of spawning.
The shades areas are where there we no data

Savage’s significant contribution was to look at the patterns of rainfall, temperature and spawning not just in the month of spawning but in each of the two months before spawning. However, he then hit a problem because the usual month of spawning was not the same throughout Britain, as his map showed, and was sometimes variable from year to year in any one region. Therefore for each record he designated the month of spawning as M0, the previous month as M1 and the month before that as M2. The reader of his book can be confused by this notation because it would have seemed more logical to denote the month before breeding as M-1 etc.

He had to use months as the unit because the meteorological records were presented as averages for a particular month. That made his conclusions for the month of spawning, M0, potentially less valuable as he explained: 

It is obvious that the weather near the end of M0 cannot affect the actions of frogs spawning at the beginning, and, for that reason, the use of a mean value for this month must introduce errors. It is, however, a property of the weather that it runs in spells, so that the Meteorological Office can head its monthly reports with a condensed summary in a short phrase, such as “Warm in the west, colder and windy in the rest of the country”. The mean values used in this work are therefore not so bad a measure as would appear at first sight.

Savage went into considerable detail in attempting to interpret the three diagrams. I will limit what follows. In general, in the month of spawning (M0), early dates for spawning are associated with higher rainfall; both relatively and high and relatively low temperatures with low rainfall are associated with later spawning. A similar conclusion can be reached about the month before spawning (M1). By contrast, two months before that of spawning (M2), temperature is far more important, higher temperature being associated with earlier spawning. An effect of rainfall is still present particularly at lower temperatures.

It is the marked association of rainfall in the preceding month or so with the date of spawning that Savage talked about on the BBC in 1959.

Savage demonstrated clearly an association between rainfall and spawning date, an association stronger at that time than the effect of temperature. It doesn’t need stating that demonstration of an association is not a demonstration of causation or that rainfall is not a proxy for some other event associated with high rainfall. However, the shift in spawning date seen after the extremely cold winter of 1947 (with ponds frozen until after day 70) argues in favour of an effect of, at least, severe changes in environmental conditions. In 1940-46 in south-west England the average day of spawning was 36 (5 February); in 1947 it was 77 (18 March).

Ecologists have often seemed reluctant to embrace experiment (a criticism that does not extend to Savage) and as far as I have been able to determine, there has been no experimental test of the effect of rainfall and temperature on the date of spawning. If, say, frogs from parts of the east of Scotland or of east Anglia (regions with late average spawning dates) were moved to Cornwall would they conform to the early spawning date observed there?

Savage thought he knew why higher rainfall is associated with early spawning and how the effect is brought about. This is an important I shall return to in a future article since it is a theme that runs through all Savage’s observations and research.

Ronald Maxwell Savage spent an enormous amount of his time—months of work he stated—in pre-computer days on extracting the data, preparing the data for analysis and doing the final statistical calculations. He thanked ‘Dr Frazer’ [John Francis Deryk Frazer, 1916-2008] for copying out the spawning dates from the Phenological Reports. Therefore, I find it odd that Deryk Frazer made so little of the approach or of Savage’s findings, confining the inadequate description to a short paragraph, in his Reptiles and Amphibians in Britain which was  published in 1983 in the New Naturalist series,. Perhaps I should not have been too surprised because in the same volume he also misinterpreted Savage’s later work.

As well as producing his diagrams, Savage, a skilled exponent of statistical analysis for reasons I will explain in a later article, used multiple regression analysis on the meteorological, location and spawning date data in order to determine the influence of individual factors. In addition to rainfall and temperature he included altitude, latitude, longitude and hours of bright sunshine. From this analysis he obtained a multiple linear correlation coefficient of 0.74, remarkably high, he suggested, for such data; I agree. That would account for just over 50% of the variation in spawning data observed and, as Savage observed, could well be an under-estimate for reasons he explained.

I thought I would try, using Savage’s multiple regression equation that he showed in his book, and average local data to see if it produced a spawning date in the right ball park. However, I soon realised that one term (temperature in M2) was missing entirely and the coefficients for latitude and longitude seemed 10x out. Using the equation as it is shown produced nonsensical answers. Unfortunately, I have found no reference to its use by others nor to its obvious errors which must have occurred during conversion of units, copying or type-setting. If it were usable it could have provided a valuable tool to test the effect of increasing environmental temperatures on spawning date (see below) and how well it corresponded to current data on climate and spawning dates. It is unfortunate that the absence of ‘hard’ editing is often evident in Savage’s papers as well as in his book. There are some obvious errors as well as the reader—well this reader—being left puzzled by what seems to be an error, omission or lack of explanation. One example is that the beta coefficients shown in the regression equation on page 143 (with one missing) are not the same as shown in Table 7 on page 145.

Savage also showed the beta coefficients obtained in his regression analysis (i.e. the slope, negative or positive, for each component). He used their relative size to estimate the main factors. Retarding influences (i.e. later spawning dates) were associated with higher temperature in M0; increase in latitude; increased sunshine in M1. Accelerating influences were: increase in longitude; raised temperature in M2; increased rainfall in M1. These conclusion can, of course, also be deduced from his diagrams.

Therefore, In any one location from year to year, the effects of geography removed and the list of influences can be shortened. Retarding are: increased sunshine in M1; high temperature in M0. Accelerating are: high temperature in M2; increased rainfall in M1. However, this could not be the whole story since, as Savage observed, although some ponds were near together the frogs spawned at different times. Again, I will return to how Savage explained the differences in a later article.

Savage appreciated the importance of an effect on winter temperature (i.e. in M2) on spawning date and its compatibility with what was known about frog physiology. Gametogenesis is temperature dependent and Trevor Beebee found that between 1979 and 1994, although spawning date did not change significantly over that period at a single site, there was a strongly negative correlation with overall winter maximum temperatures*.

It would seem that the tradition of making rather little of Savage’s extensive work and statistical analysis, as exemplified by Frazer’s book and pointed out by Trevor Beebee, lives on. For example, in a recent paper on the possible effects of climate change on breeding in the Common Frog, Savage is mentioned but only in respect of noting that spawning dates may be different in ponds in the same area. Completely ignored in this new phenological analysis across a number of sites from 1994 onwards was the importance of the timing of rainfall in the months before spawning as found by Savage and the acceleration of spawning by higher temperatures in M2 but a retardation of spawning with high temperatures in M0. There was incidentally, no statistically significant trend for spawning dates to be earlier, although there was a tendency in that direction. The plot of spawning day versus year would have been dismissed by a late colleague as just ‘a swarm of bees’. Savage may have been wrong—or right—in his conclusions but his findings have to be explained in the light of further evidence, not ignored, as well as being taken into account in designing statistical procedures to analyse changes with time.

Finally, and with more on Ronald Maxwell Savage to come, the original data sources still exist. Somebody could re-extract the information, which would be time consuming but the months spent on statistical analysis—which could go much further than even Savage could crank out by hand—would be reduced to minutes. Given the importance of determining the effect of anthropogenic warming, these historical phenological records could be of much greater significance than Savage could ever have anticipated.

Beebee TJC. 1995. Amphibian breeding and climate. Nature 374, 219–220. 

Savage RM. 1935. The influence of external factors on the spawning date and migration of the Common Frog, Rana temporaria temporaria Linn. Proceedings of the Zoological Society of London 105, 49-98.

Savage RM. 1961. The Ecology and Life History of the Common Frog. London: Pitman.


Tuesday 16 March 2021

The Mouse Adrenal X-Zone Revisited


This photomicrograph (from here) shows the cross section of a female mouse.
The medulla (M), X-zone (XZ) and zona fasciculata (ZF) of the cortex are labelled.
The outer, zona glomerulosa, of the cortex can be seen but is not labelled 


For a few months in 1965-66 I worked on the aptly-named X-zone of the mouse adrenal in Hong Kong. In 2018 I wondered what had happened to research on the X-zone. Had its function been discovered? The answers were in fact ‘rather little’ and ‘no’ and so I offered to give a talk at an annual Society for Endocrinology meeting in an attempt to stir up some interest in a problem that has intrigued those interested in the workings of the adrenal gland since the 1920s. This I did in November 2019 at Brighton meeting of the British Endocrine Societies.

I have written below an account of the cellular origins of the X-zone because it now transpires that we were misled in the 1960s by the misinterpretation or misreporting of early findings and that, in fact, the early workers were right about where it came from during development of the embryo. In addition, nobody it seems had spotted an important paper published in 1942 that made our experiments in Hong Kong unnecessary. But first the personal historical background:

Four weeks after we had arrived in Hong Kong, I received an aerogramme from John Phillips he had written on 28 November 1965. He was on his first ‘long leave’ from the University of Sheffield and was spending it back in Sheffield with Ian Chester Jones, where, until December 1962 he had been a lecturer. We had been up to Sheffield several times before leaving for Hong Kong on 1 November and the general idea was that I should have a look to see if steroid hormones found in vertebrates occur in invertebrates as well. I was beginning to see what we had in terms of chromatographic equipment and chemicals in order to make a start when that aerogramme arrived. He wrote:

…I listened to Prof Paul Delost give a lecture on the X Zone last night and I was surprised to hear that he considers the X zone to be under medullary control. He bases this conclusion on the absence of an X zone in an adrenal in which the medullary tissue has been aspirated from the centre of the gland with a needle and vacuum pump—the other gland remaining as a control. The interesting thing about this preparation is that if you castrate the post-pubertal male the X zone reappears in the gland with a medulla but not in the other adrenal without a medulla…But the main criticism of Delost’s approach is that he destroys the vascular bed of the adrenal. This can be overcome by an operation called “enucleation” in which the whole of the adrenal is expressed leaving only the capsule from which a new adrenal cortex regenerates. Will you get some male mice and enucleate adrenals before puberty…

That I did and the results were clear. Given though the techniques available at the time it was difficult to envisage taking the approach further. The caravan moved on.

Cellular Origins of the X-Zone

After its description1 but misidentification in 1924 by Kiyoshi Masui and Yasushige Tamura of the Imperial University of Tokyo (with a further publication in English in 19262) and  its naming in 1927 to reflect its unknown function by Evelyn Howard (1904-1999) then at Stanford3, the X-zone of the mouse adrenal excited the interest of pioneering endocrinologists who did not yet describe themselves as such. The structure, which develops after birth between the cortex proper and the medulla, still befits its name; the function of the X-zone remains unknown4,5. Early work was concerned with its origins and with its hormonal control since it disappears at puberty in males and during first pregnancy in females6.

In terms of hormones in the circulation, androgens cause the X zone’s disappearance while LH (luteinising hormone) from the pituitary is necessary for its maintenance4,7. Although excellent research was done in the early years on the possible cellular origins of the X-zone it is only more recently that cell lineage studies using molecular markers have provided further evidence that it is derived from the fetal or inner adrenal cortex, as opposed to the definitive or outer cortex which forms the well-known zones of the adrenal: glomerulosa, fasciculata, reticularis (references in4). Therefore, the X-zone of the mouse appears to be homologous with the human fetal adrenal cortex, which, as its names implies is present only in the developing fetus. Such an origin was suspected by some early workers who referred to the ‘human fetal X-zone’ but strongly denied by others. Other possible homologues are the juxtamedullary zones of various size and appearance observed in some other eutherian mammals (cat, rabbit, voles, hamsters, and shrews)6,8.

In this short article I first consider whether the more modern findings on the origins of the X-zone are consistent with the early studies since the over-riding impression created in the mind by reading reviews and papers from the latter half of the 20th century is that the X-zone is derived from the inner cells of the zona fasciculata or, in other words, is just another zone of the definitive cortex. Then I review the evidence from little-known perturbative experiments, published, to modern eyes, in obscure places, that throw light on the origins of the X-zone.

That the X-zone is derived from the fetal or inner cortex is entirely consistent with the findings of Harry Waring9 who was then working in Liverpool for an M.Sc. The topic was suggested to him by a forgotten promoter of endocrinology in Britain, a famed lecturer in zoology, Ruth Culshaw Bamber (1889-1970) who was always known as Mrs Bisbee.  Horace ‘Harry’ Waring (1910-1980) showed in 1935 that during embryonic development there is—initially—an intermingling of the cortical and medullary elements. Remodelling then concentrates the medulla until there is a clear separation from the cortex. He identified cells, comprising what he called the interlocking zone, between the medulla and cortical elements. These cells became concentrated around the time of birth into a layer around four cells thick. The outer cortex in the meantime was growing and forming the usual zones. But it was that layer of interlocking cells that went on to form the X-zone after birth. Later, as the X-zone degenerated there was left a ‘medullary connective tissue capsule’ or ‘juxtamedullary capsule’ around the medulla formed, it was presumed, from the collapsed stroma9.

In 1928, Ruth Deanesly (1901-1997) in London had already observed that after degeneration of the X-zone some of its cells remained around the juxtamedullary capsule10. Therefore, the key early finding that a secondary X-zone forms after castration of male mice can be explained by growth from these cells, i.e. remnants, capable of division, of an inner (fetal) cortex rather than from a differentiation of cells from the inner zone (z. fasciculata in the mouse) of the outer (definitive) cortex.

I cannot explain why the view prevailed, despite evidence to the contrary, that the X-zone was part of and derived from the outer or definitive cortex. As one example, the following is from the highly influential review written by Helen Wendler Deane (1917-1966) published in 19628, four years before her early death:

These [X-zone] cells differentiate postnatally, at about 2 weeks, from the inner portion of the fasciculata (Whitehead 1933a, Waring 1935).

The problem with this statement is that neither Raymond Whitehead11, working in Manchester, nor Harry Waring9 drew any such conclusion. Only Waring of the two studied the origins of the X-zone and his conclusion was, as I have noted above, entirely different.

All the above evidence, even the sophisticated and relatively recent cell lineage studies, have been observational. Those seeking direct, experimental evidence that throws light on the origin of the X-zone would at first sight be discouraged since it all appeared over 50 years ago in obscure publications and/or written in French while one important paper had, I discovered recently, been overlooked entirely.

Until I found that paper, the first experimental work that has a bearing on the problem was thought to have been that in the 1960s from Paul Delost’s laboratory at the University of Clermont-Ferrand in France and in particular that of his student Parviz Chirvan-Nia who later returned to Tehran University of Medical Sciences in Iran. They took advantage of the fact that a secondary X-zone develops after castration in male mice. Their most pertinent finding12,13 was obtained in mice from which the adrenal medulla had been removed by aspiration through a very fine pipette, a technically difficult procedure. In males in which the entire medulla had been removed, a secondary X-zone failed to develop after castration. By contrast, if even a small piece of medulla remained, an X-zone developed around it. The effect was local; the untreated contralateral adrenal was unaffected.

All of Delost’s work from the 1950s onwards was published in French and had received little attention in the English-speaking world. However, on 27 November 1965 Delost was invited by Ian Chester Jones (1916-1996) to give a seminar in Sheffield, having been one of Chirvan-Nia’s external examiners earlier that year. His old student, John Guest Phillips (1933-1987; FRS 1981), on leave from the University of Hong Kong, was also there and while intrigued by Delost and Chirvan-Nia’s work he and Chester Jones were concerned that, in aspirating the medulla, the venous drainage from the adrenal would have been destroyed. They thought that a complementary approach, that of enucleation, in which end of the adrenal is snipped off and the medulla and most of the cortex squeezed out, would be a useful test. After enucleation the outer cortex redevelops and shows a normal zonation but without a medulla. As a result, John Phillips sent me an aerogramme the next day asking if I, having arrived in Hong Kong four weeks earlier as a PhD student, would take it on. This I did and the results were identical to those obtained by Chirvan-Nia and Delost: a secondary X-zone failed to develop if the medulla had been removed completely; if even a small remnant of medulla remained after incomplete enucleation, an X-zone developed around it14.

In 2019 while preparing a talk5 for the Society of Endocrinology on what has happened to research on the X-zone I discovered our experiments had simply confirmed what had already been published. The same approach, enucleation, had yielded the same results in 1942. The paper, which appears not to have been quoted by anybody working in or reviewing the field, was by Murchie Kilburn McPhail (1907-1987) and his student H.C. Read, of Dalhousie University in Canada15. I can only assume that it was missed because it appeared in wartime, albeit in a leading journal, when scientists were otherwise occupied. However, another paper from the same authors from the same year was picked up and referred to.

Suggestions as to possible mechanisms as a result of these experimental approaches, for example what would now be termed a paracrine effect of medullary cells on the inner cells of the  cortex against the direction of blood flow, can now be discarded since the removal of the medulla would by any technique result in the extirpation of the inner cortical anlagen formed in and around the juxtamedullary connective tissue capsule after degeneration of the X-zone.

In conclusion, the original observations by Harry Waring on the origins of the X-zone and the experimental evidence on the necessity of the adrenal medulla for the presence of an X-zone are entirely consonant with the X-zone being derived from the fetal or inner cortex and not the definitive or outer adrenal cortex. The recognition that there are two populations of cortical cells only one of which is responsible for the classical zonation of the adult adrenal gland explains many of the false leads followed, blind alleys entered and bold assertions made by those working in the field during the middle decades of the 20th century.

Research over the past 96 years has established the hormonal control of the adrenal X-zone and seemingly settled its cellular origins. Will we also know its function by the time of the centenary of its discovery?

1. Masui K, Yamura Y. 1924. The effects of gonadectomy on the structure of the suprarenal glands of mice, with special reference to the functional relation between this gland and the sex gland of the female (Translated). Nihon Chikusan Gakkaiho 1, 55–79.

2. Tamura Y. 1926. Structural changes in the suprarenal gland of the mouse during pregnancy. Journal of Experimental Biology 4, 81–92.

3. Howard E. 1927. A transitory zone in the adrenal cortex which shows age and sex relationships. American Journal of Anatomy 40, 251-293.

4. Huang C-C J, Kang Y. 2019 The transient cortical zone in the adrenal gland: the mystery of the adrenal X-zone. Journal of Endocrinology 241, R51–R63.

5. Peaker M. 2019. What happened to the adrenal X-zone. Endocrine Abstracts 65 SE 1.1. DOI: 10.1530/endoabs.65.SE1.1.

6. Chester Jones I. 1957. The Adrenal Cortex. Cambridge: Cambridge University Press.

7. Gannon A-L, O’Hara L, Mason JI, Jørgensen A, Frederiksen H, Milne L, Smith S, Mitchell RT, Smith LB. 2019. Androgen receptor signalling in the male adrenal facilitates X-zone regression, cell turnover and protects against adrenal degeneration during ageing. Scientific Reports 9, 10457.

8. Deane, H.W. 1962. The anatomy, chemistry, and physiology of adrenocortical tissue. In The Adrenocortical Hormones Part 1. Handbuch der Experimentellen Pharmakologie, edited by Eichler O & Farah A, subedited by Deane HW, 1-185. Berlin, Springer.

9. Waring H. 1935. The development of the adrenal gland of the mouse. Quarterly Journal of Microscopical Science 78, 329–366.

10. Deanesley R. 1928. A study of the adrenal cortex in the mouse and its relation to the gonads. Proceedings of the Royal Society B 103, 523–536.

11. Whitehead R. 1933. The involution of the transitory cortex of the mouse suprarenal. Journal of Anatomy 67, 387–392.

12. Chirvan-Nia P. 1965. Données nouvelles sar la zone X surrénalienne de las souris. Doctoral Thesis, Université de Clermont.

Brudieux R., Chirvan-Nia P., Delost P. 1966. Sur les relations directes entre la médullo-surrénale et le cortex surrénal. Journal de Physiologie, Paris 58, 213-217.

14. Peaker M, Phillips JG,. Peaker SJ. 1967. A relationship between the medulla and the X-zone of the mouse adrenal. In Proceedings, Third Asia and Oceania Congress of Endocrinology, edited by Litonjua A, vol. 2 317–321. Manila.

15. McPhail MK, Read HC.  1942. Regeneration of adrenal gland following enucleation and transplantation with special reference to X-zone. Endocrinology 31, 486–492.