Brazil Nut Effect ~ Exploring the Glen Rose Limestone, Part 2

My first impression wasn’t wrong, it was just incomplete because I’d fallen victim to the Brazil nut effect.

In my previous post, I described the shells of Orbitolina texana, an agglutinated foraminifera, which appear in staggering numbers in a sample of material collected in Texas from the Glen Rose Limestone.  As far as I could see, this material, laid down in the Lower Cretaceous about 113 to 110 million years ago, consists almost completely of these foram shells.  For this present post, I took a closer look.

I washed the sample and then worked it through two sieves, creating two portions that differed by the size of their component pieces:  the first with anything larger than 1 mm, and the second with anything equal to, or smaller than, 1 mm and larger than 0.177 mm (177 microns).  I should add that any very large object, such as the example of a Salenia texana (a type of sea urchin) shown below, was removed in advance from the first portion.  (This fossil is about 2 cm across and stands about 1 cm tall.)

From each sieve, I shook material into tall envelopes secured with clips.  (In a prior post, I considered some aspect of the effects on microfossil searching of the mesh size of sieves.)

For this post, I’m not interested in that first portion (greater than 1 mm) which, under the microscope or with the naked eye, for that matter, basically confirms what one could learn just by looking at the ground where this material was collected  (the previous post includes a photograph of material in situ):  aside from some fairly poorly preserved larger invertebrate fossils, the matrix appears to be composed almost exclusively of Orbitolina texana shells

It’s when I began to search through the second portion (material between 0.177 mm and 1 mm in size) that things got interesting (well, at least, they did for me).

For awhile, what I shook from the second-portion envelope into the sorting tray seemed little different from what I’d seen in the first portion with the larger specimens.  There were many beautiful, albeit small, O. texana shells.  These, I believe, are from juveniles.  Scattered amid the O. texana shells were some fragments of sea urchin spines and pieces of urchin plates with tubercles (where spines were attached).

So, early in the process of working down through this second envelope, the picture of the Glen Rose Limestone at this location that I’d already formed was unchallenged.  I remained convinced that this fauna was thoroughly dominated by a single species of foram.

But, then, as the material from lower in the envelope tumbled into the sorting tray, my perspective changed, the picture became more complex.

There had been a more diverse population at this site than I’d imagined.  Fossil shells from ostracodes began to appear frequently, now dominating the sorting tray.  Equally impressive, these ostracode shells came from a rich array of species.  Pictured below are several examples of Eocytheropteran trinitiensis (first picture), Cythereis ornata (middle picture - note, the specimen at the top center is a carapace (i.e., having both valves) and is resting on its dorsal side), and Paracypris weatherfordensis (bottom picture).

Identification of these specimens is based on H.C. Vanderpool, Fossils From the Trinity Group (Lower Comanchean), Journal of Paleontology, Volume 2, Number 2, June, 1928, and Frederick M. Swain and Philip M. Brown, Cretaceous Ostracoda From Wells in the Southeastern United States, North Carolina Department of Conservation and Development, Bulletin Number 78, 1964 (this latter piece revises many of the ostracode name assigned by Vanderpool).  For these references and others in this post, I provide links only to publications that do not reside behind paywalls.

Vanderpool identified nine different species of ostracode that he found in the two Glen Rose locations he sampled.  He characterized five of these as "abundant."  Swain and Brown revised several of Vanderpool's identifications, including one in which they concluded he had considered adult and juvenile forms of one species as two separate ones.

This is an aside, but I have to note that, in light of what I've found in my Glen Rose sample, it's quite stunning to read Vanderpool's assessment of the microfauna in the environment in which the Trinity Group formations (which include the Glen Rose Limestone) were laid down.  It suggests the difficulty in basing such an assessment on merely a handful of sites (six in his case, two of which featured Glen Rose Limestone, or one in my case).  From what he found, he concluded that the near-shore environment of these locations posed such challenges that only "the more hardy forms" of microfauna could handle it, primary among them were, in his estimation, the ostracodes.  In contrast, foraminifera, he concluded, appeared only "sparingly."  Yes, I'll grant him that ostracodes do turn up in the Glen Rose in large numbers, but "sparingly" is not the adverb for the O. texana at the site of my sample.

Though I don’t know if most of the ostracodes I was finding co-existed with the hordes of O. texana forams, I suspect they did.  Somehow these ostracodes in the substrate earned their living while navigating around the O. texana.

As my sense of the Glen Rose microfauna changed (coming closer, perhaps, to the reality of this formation at this time and in this place), I realized that this modification in perspective tracked with how far down I’d gone into the container with that second portion of material.  Indeed, the material toward the bottom of this envelope was markedly smaller than that at the top.

And, so, the Brazil nut effect had come into play.

A good place to start is with breakfast.  Most mornings, my breakfast includes some Chex Granola Mix which I most enjoy when a new package is first opened.  It’s then that the big “honey nut clusters” and whole Chex squares tumble into my bowl.  Several mornings later, as I near the end of the bag, I’m down to bits of nuts, a bunch of oat and rice flakes, and fragments of Chex squares – not so enticing.  The view below, looking down into a newly opened package, shows the big stuff gathered at the top.

This particle segregation is popularly known as the Brazil nut effect because it is so evident in containers of mixed nuts where the Brazil nuts (or other large nuts) gather at the top.  (On occasion, it's been called the muesli effect.)  Rather than being some trivial, though interesting, topic, the Brazil nut effect is critical for a range of industries, such as food processing or pharmaceuticals, in which controlling how dry mixtures behave is important.  It’s also a key context within which microfossils are hunted.

This much-studied phenomenon continues to confound.  The original hypothesis was that, in containers with a mixture of objects (some larger than others), shaking causes the larger ones to percolate to the top.  (These larger objects are called intruders in the literature.)  Under this reasoning, it was assumed that, in a shaken container, spaces created beneath the intruders fill with smaller members of the mixture.  With more shaking, this filling process ultimately boosts the intruders to the top.   Once there, their sheer size precludes them shifting down.

However, further research concluded that, instead of percolation, convection is predominantly at work with an upwelling in the middle of the mixture (pulling up everything, not just the intruders) and a downwelling at the outer margins.  The problem for the intruders is that, once they’ve ridden the upwelling to the top, they get stuck; they’re just too big to descend with the margins.

Unfortunately, for my sense of clarity, the story doesn’t end there.  Apparently, same-sized light and heavy intruders can behave differently, challenging both the percolation and convection explanations.  In this reverse Brazil nut effect, the lighter intruders, unlike their same-sized heavier counterparts, can work their way down to the bottom of the container.  But, wait, there’s more.  Density and size of elements in the mixtures can interact in complex ways.  For instance, holding the density constant, a size threshold comes into play with those objects below it falling and those above rising.  If, instead, the size of intruders is held constant and density is allowed to vary, it appears that intruders at either extreme – high density or low density – will rise faster than those intruders whose density is between the extremes.  If that weren’t enough, it seems that the frequency of vibrations applied to the mixture makes a difference as does the ambient air pressure.

After penning a concise overview of the state of understanding of this phenomenon (and its counterpart), biomedical engineer Troy Shinbrot, who co-authored the paper that first identified the reverse Brazil nut effect, couldn’t resist this wonderful quotation, which he attributed to Mark Twain (The Brazil Nut Effect – In Reverse, Nature, Volume 429, May 27, 2004, p. 352-353):

 The researches of many commentators have already thrown much darkness on this subject, and it is probable that, if they continue, we shall soon know nothing at all about it.

Note [added later]:

I did test whether the Brazil nut effect was actually in effect with my portions of Glen Rose material, but these were a bit indirect.  I transferred some of the first portion (greater than 1 mm) material into a glass jar and gently shook it. Most of the largest pieces in the jar slowly gathered at the top.  I attempted to replicate what shaking might do to the second portion (between 0.177 mm and 1 mm) by adding a few slightly larger O. texana shells to material from that portion that I had already gone through in my hunt for microfossils.  (I simply could not bring myself to add back the ostracodes I had found and see if they ended up lower in the container.)  These larger O. texana shells, as expected, popped to the surface after only a few shakes.  Conclusive evidence?  No, but certainly supportive.


Addendum:  Dating the Glen Rose Limestone

As far as I can tell, the Glen Rose Limestone, part of the Trinity Group, dates from close to the end of the Aptian through much of the lower Albian Ages (from roughly 113 to about 106 million years).  The sample I’m working with probably is from the Lower Glen Rose (about 113 to 110 million years ago).  I crudely estimated these dates from Figure 1 in the paper by Ernest A. Mancini and Robert W. Scott, titled Sequence Stratigraphy of Comanchean Cretaceous Outcrop Strata of Northeast and South-Central Texas:  Implications for Enhanced Petroleum Exploration (Gulf Coast Association of Geological Societies Transactions, Volume 56, 2006, p. 541).  As for the idea that this material had to be from the Lower Glen Rose, I have relied on Raymond Douglass' assertion that O. texana is only found in that part of the Glen Rose (The Foraminiferal Genus Orbitolina in North America, U.S. Geological Survey, 1960, p. 6).