Water hardness, stream size, and A.E. Boycott: A New River reminiscence

Just across the Blue Ridge, where the high meadows lay,

And the galax spreads through the new mown hay,

There's a rusty iron bridge, 'cross a shady ravine

Where the hard road ends and turns to clay [1].

The New River is the second-oldest river in the world, just a bit younger than the Nile.  I have heard that assertion stated so often and so forcefully that it must be so.  Born on the slopes of North Carolina’s Grandfather Mountain (G, down below), the New River first flows rather improbably to the northeast, through high meadows parallel to the Blue Ridge, into the Commonwealth of Virginia.  Then, quite astoundingly, the river shifts its course northwest, orthogonal to the Appalachian Mountains, near the little city of Blacksburg (B), and cuts a deep notch through the Allegheny Plateau, diagonally across West Virginia to The Ohio.  The earliest explorers of the American interior named its lower half the Kanawha River, never imagining that the New River and the Kanawha River might connect through those hundreds of miles of rugged terrain.

New River at Grandview

My mother was born and raised in the little town of Floyd, Virginia, perched on the New River plateau, looking down over the Blue Ridge (M).  And my father was born and raised in Rock Castle Gorge, deep in the Blue Ridge down below (F).  Many were the sparkling summers I spent rocking on my grandmother’s porch, many were the spring times I fished the chilly creeks for trout, many the falls I hunted the dark forests for squirrels.

 

With a suitcase in his hand

There the lonesome boy stands

Gazing at the river sliding by beneath his feet,

But the dark water springs from the black rocks and flows

Out of sight where the twisted laurel grows.

 

So, I matriculated at Virginia Tech, the University of Blacksburg, in the fall of 1973.  And I have shared with this readership quite a few anecdotes about my education in those hallowed halls [2], especially highlighting the impact of my mentor, Dr. E. F. (Fred) Benfield.  Longtime readers might remember that, even as I was finding a seat in my freshman biology class, plans were well advanced to construct a double-dam pump-storage facility on the upper New River at the Virginia / North Carolina border, which would have sunk 100 miles of the world’s second-oldest river into inky blackness and inundated thousands of square miles of lovely farmland.  And the Virginia Tech Center for Environmental Studies had been contracted to prepare the Environmental Impact Statement [3].

 

And so it was that I, a mere sophomore of age 19, found myself checking out boots, buckets and nets from the storeroom of the Virginia Tech Center for Environmental Studies, pulling the keys to a pickup truck from the pegboard, and driving off into the high meadows of the upper New River for my first systematic survey of a freshwater molluscan fauna [4].  And in my undergraduate research thesis, defended in May of 1977, I reported a modest 6 species of unionid mussels, 4 species of pisiidid clams, 4 species of prosobranch snails, and 6 species of pulmonate snails from 87 sites sampled across the drainage in ten counties of southern Virginia and northwestern North Carolina [5].

 

It materialized that those 20 species of freshwater mollusks were not all evenly distributed across my ten-county study area.  Even to my young and untrained eye, patterns manifest themselves.  And the most striking pattern was closely correlated with the underlying geology of the upper New River basin.  Which sent me to the library, looking for clues.

The geology of the New River Valley of Virginia

I do not remember the day or the hour I first discovered the wonderful 1936 work of A. E. Boycott [6], but I do remember the impact.  Boycott’s 70-page review was a thing of beauty; so complete, so rigorous, so bubbling over with cheerful facts about the biology of the creatures we both obviously loved that 20 years later, I patterned an entire book after it [7].  And in retrospect, may have produced but a pale imitation.

 

So Boycott, after reviewing the general biology and local habitats of the 62 species of freshwater mollusks inhabiting the Island of Great Britain, turned to “The relevant qualities of the habitats.”  Item (b) in his list of eight relevant qualities was “lime,” and item (c)  was “reaction,” by which he meant pH, which (he hastened to point out) was essentially the same thing as lime [8].

 

Boycott went on to classify (and map) the 62 British species as 25 “soft-water” species (“all those which we can find in soft water without surprise”) and 30 species “calciphile or calcicole in the sense that they are habitually found in hard water [9].”  He could not identify any British species as “calcifuge,” i.e., restricted to soft water.  Thus, overall molluscan diversity increased in the hard waters, as the 30 calciphiles were added to the 25 softwater species in the richer environments.

 

Well, that explains a lot right here in the good old USA, I thought to myself.  From its North Carolina origin through the first (roughly) hundred miles of its journey, the New River runs through ancient metamorphic rocks such as gneiss and schist, remaining quite soft.  But about halfway to its hard left turn at Blacksburg, the New enters a region of limestone and dolomite, and the water hardens up.  The distribution of most of the unionids and a couple of the gastropods (Pleurocera shenandoa, Physa gyrina) seemed to be restricted below that invisible barrier, as though they were, in Boycott’s terminology, “calciphile.”

 

But there was a second obvious factor in the distribution of the upper New River mollusk fauna, and Boycott had that one covered as well.  Item (d) of Boycott’s eight “relevant qualities” was “Size and Volume.  The larger units of water are liable to contain the more Mollusca.”  And indeed, most of the unionid mussel species were confined to the main New River, apparently unable to inhabit the smaller tributaries.  That also seemed true for Campeloma decisum.

Dillon & Benfield [16]

That’s pretty much where I dropped the shovel for my 1977 undergraduate thesis and sat down in front of the typewriter.  Each of my four subheadings under Results and Discussion – Unionidae, Sphaeriidae, Prosobranch Gastropods, and Pulmonate Gastropods – had a section entitled “Effect of Hardness” and a section entitled “Effect of Stream Size.”  All four subheadings also had a section entitled “Effect of perturbation” or “Effect of (artificial) enrichment.”  This was the 1970s, after all, I had to get pollution in there somehow.  I defended in May, got married in June, and moved to Philadelphia in July.

 

Ah, but.  Hidden deep inside my thesis was the germ of an idea.  Under the “Unionidae” subheading was also a brief section entitled “Interaction of factors.”  And there I speculated, at the age of 21, “Perhaps hardness and stream size interact in some manner so that a large stream can support Elliptio dilatata even though its hardness may be low, and a small stream can support mussels if it has high hardness.”  That hardness x stream size interaction was also obvious in the pulmonate snails.  About the origin of the phenomenon, at my young age, I would not hazard a guess.

 

Past the coal-tipple towns in the cold December rain

Into Charleston runs the New River train,

Where the hillsides are brown, and the broad valleys stained

By a hundred thousand lives of work and pain.

 

Ecology at The University of Pennsylvania and ecology at Virginia Tech are as different as Philadelphia and Blacksburg.  At Penn, I found the intellectual focus entirely upon the interactions among organisms, not the interactions between organisms and their environment.  Density dependence was the key, density-independence a quaint anachronism.  I remember vividly the argument made by the chairman of my graduate committee, Bob Ricklefs.  “A population without density-dependent control will either go extinct, or cover the world ass-deep.”  Since freshwater gastropods exist [10], and we are not ass-deep in them, they must be under density-dependent control.  Food availability or predation might certainly qualify as potential controls for their distribution, perhaps, certainly not the availability of calcium to build their shells.

 

Robert MacArthur’s theory of island biogeography was also still very much en vogue at Penn in the late 1970s, with its focus on island size.  And somewhere during my first year of graduate training, it dawned on me that both the size of a stream and the hardness of its water might affect its productivity.  And that the quantity of food might be controlling freshwater mollusk distributions in the upper New River, neither the calcium nor the stream size directly.

 

And so, at age 22, I hazarded a guess.  In July of 1978 I drove back down I-95/I-81 south from Philadelphia into those high meadows of the Upper New River where I had spent the summers of my youth, with beakers, cylinders, burettes, bottles of indicators and a 0.02 N sulfuric acid titrant carefully stowed behind the hatch of my 1973 Pinto.  I revisited and re-sampled every one of those 87 sites I had collected in the years previous, this time taking a streamside measurement of alkalinity, more reliable than pH, much easier to measure than calcium or overall hardness.  And this time I estimated a rank abundance for each of the five New River pulmonate snails (excluding limpets, which were omnipresent): Physa acuta, Physa gyrina, Lymnaea humilis, Lymnaea columella, and Helisoma anceps [11].

Dillon & Benfield [16]

I found pulmonate snails at 26 of the 87 sites, marked letters A – Z in the map above.  Confirming and deepening my undergraduate results, pulmonates were common and widespread in the main New River and most of its tributaries downstream from where the river entered the limestone/dolomite zone.  Above that zone, however, pulmonates were generally found only in the main river itself.

 

I summed the rank abundances of the five species into an overall measure of pulmonate abundance for each site.  And I calculated the Kendall rank correlation between pulmonate abundance, alkalinity, stream drainage area and (here’s the key!) the alkalinity x drainage area interaction.  The correlation with interaction (0.35**) was greater than either alkalinity or drainage area alone.

 

OK, one last thing.  Notice in the table above that alkalinity and drainage area were negatively correlated.  I guess that’s not too surprising – a little creek running through a limestone valley can get much harder than a big river, buffered as it is by its large catchment.  This gave me the idea of a nonparametric partial correlation coefficient, analogous to (parametric) partial correlation – a correlation between two variables holding a third variable constant.

Dillon & Benfield [16]

So, I invented nonparametric partial correlation [14].  The table above shows that the Kendall rank correlation between abundance and interaction remains high, even if alkalinity is partialled out (0.31) or if drainage area is partialled out (0.27).  The primary phenomenon is the interaction – the increased productivity that both water hardness and stream size promote – not the calcium nor the stream size directly.

 

When the paper by Dillon & Benfield [16] finally reached publication in 1982, I honestly thought I would become famous.  Brilliant young scientist invents novel statistical technique to answer a fundamental question of freshwater biology!  Alas, no.

 

In a tar-paper shack out of town across the track

Stands an old used-up man trying to call something back

But his old memories fade like the city in the haze

And his days have flowed together like the rain

 

And the dark water springs from the black rocks and flows

Out of sight where the twisted laurel grows

 

Notes:

 

[1] The lovely and haunting lyrics interleaved with this month’s essay come from a song entitled “Twisted Laurel” by The Red Clay Ramblers, one of the greatest bands of the postmodern era.  Some good music is still being made today, but for mysterious reasons has fallen out of fashion with the popular multitude.  Exactly the same could be said for science.

 

[2] A small sample of previous essays in which I have reminisced about my undergraduate experiences at Virginia Tech and traced their subsequent influence on my scientific career:

• To Identify a Physa, 1975 [6May14]
• Pleurocera shenandoa n.sp. [11Mar19]
• Interpopulation gene flow: King Arthur’s lesson [7Sept21]
• Growing up with periwinkles [6Apr23]

[3] For more about APCO’s Blue Ridge Project and its ultimate fate, see:

• Woodard, R. S., Jr. (2006) The Appalachian Power Company along the New River: The defeat of the Blue Ridge Project in historical perspective.  M.A. Thesis, Virginia Tech, Blacksburg.  139 pp. [pdf]

 [4] We never change.  Fifty years later, I am still doing exactly the same thing.

 

[5] Dillon, R.T., Jr (1977) Factors in the distributional ecology of upper New River mollusks (Va/NC).  Undergraduate research thesis, Virginia Tech. [pdf]

 

[6] Boycott, A.E. (1936) The habitats of freshwater molluscs in Britain.  Journal of Animal Ecology 5: 118 – 186.

 

[7] Dillon, R.T., Jr. (2000) The Ecology of Freshwater Molluscs. Cambridge University Press.  509 pp.

 

[8] Boycott verbatim: “Broadly speaking, for the natural waters of this country the reactions run parallel with the quantities of calcium.”

 

[9] Boycott did not categorize 7 of the 62 British species because he felt that his data were insufficient.

 

[10] OK, I realize that I have touched a controversial point.  The vast majority of my colleagues today, including (quite likely) most of the handful of you who will ever read this footnote, believe, as an article of faith, that freshwater gastropod populations are indeed going extinct.  Possibly that every living creature on this earth, except cockroaches, mosquitos, thee and me, is going extinct?  And I will grant you all that the vast majority of all species that have ever lived have, indeed, gone extinct.  That is evolution.  That is not a crisis; that is not even bad.  That is normal.

 

[11] Back in 1978, I was still using George Te’s [12] identifications for the Physa, so Physa acuta = “hendersoni” and Physa gyrina = “pomilia.”  I identified the Lymnaea humilis as “Fossaria obrussa” [13] and used “Pseudosuccinea” as the genus for columella.  Science advances.

 

[12] For the complete story, see:

• To identify a Physa, 1971 [8Apr14]
• To identify a Physa, 1975 [6May14]
• To identify a Physa, 1978 [12June14]
• To identify a Physa, 1989 [3Oct18]
• To identify a Physa, 2000 [6Dec18]

[13] For the complete story, see:

• The American Galba and The French Connection [7June21]
• The American Galba: Sex, Wrecks, and Multiplex [22June21]
• Exactly 3ish American Galba [6July21]

[14] I was unable, however, to offer any statistical inference on my newly-invented Kendall partial rank correlation coefficients.  I was experimenting with Monte Carlo techniques at the time [15], and (in retrospect) should have done so.   Then I would have become famous.  Surely.

 

[15] Dillon, R.T., Jr. (1981) Patterns in the morphology and distribution of gastropods in Oneida Lake, New York, detected using computer-generated null hypotheses. American Naturalist 118: 83-101.  [pdf]

 

[16] Dillon, R.T. and E. F. Benfield (1982) Distribution of pulmonate snails in the New River of Virginia and North Carolina, U.S.A.: Interaction between alkalinity and stream drainage area. Freshwater Biology 12: 179-186. [pdf]