The Forest Trapped in Resin

AAmber is not a mineral but a fossil in its own right — tree resin that has shed its volatile fractions and polymerised into a solid capable of surviving a hundred million years of burial. Three deposits dominate the literature: Baltic amber, laid down roughly forty-four million years ago; Dominican amber, at about sixteen million years the youngest of the three; and the amber of Kachin State in northern Myanmar, dated by uranium-lead methods to 98.8 million years and by some distance the richest, with more than two thousand insect species described from it since 2010 alone. What draws palaeontologists is fidelity. A midge caught in resin keeps its wing hairs, its gut contents and occasionally its colour, while the same animal in a lake shale survives as a stain. Yet fidelity of a specimen is not the same as fidelity of a sample, and whether amber offers a census of an ancient forest or a peculiar advertisement for one has divided the field for thirty years.

BIngrid Sørholt, a taphonomist at the Uppsala Institute for Fossil Preservation, holds that the sample is the problem and that no amount of optical brilliance will repair it. Between 2011 and 2022 her group re-examined 12,400 inclusions from three deposits, measuring every specimen and assigning it to a habit class. Eighty-eight per cent were under five millimetres long; animals that lived on trunks or flew close to them made up 71 per cent of the total, against 19 per cent in compression fossils from lake beds of the same age and region. Ground-dwelling and aquatic groups, abundant in the shales, were almost absent. Over the same period the number of amber species formally described each year roughly tripled, an expansion that reads in review articles as a widening picture of Cretaceous life. What did not widen was the range of habits represented. The trouble, on her account, lies in what resin can physically reach rather than in how much amber has been sawn open.

CThe account yields a prediction, and Sørholt has tested it. If the resin rather than the forest sets the sample, two ambers from different continents ought to resemble each other more closely than either resembles the lake shale laid down beside it. Across nine deposits her group found exactly that, with an average similarity of 0.62 between ambers and 0.24 between an amber and its neighbouring shale. The result was confined to resin produced by araucarian conifers; where the source tree was a flowering plant, the amber tracked its local shale almost as closely as it tracked other ambers. Sørholt therefore treats the trap as a filter whose settings shift with the tree, not as a uniform distortion that one correction factor could remove. Where the source plant cannot be identified from the resin's chemistry — the case at perhaps a third of sites — she accepts that no correction can be attempted at all.

DTeodor Vaslin, who works on fossil behaviour at the Institute of Palaeoentomology in Cluj, answers that a biased sample is still the only sample that shows animals doing anything. Resin engulfs a scene in minutes, and the record therefore holds interactions that no other kind of preservation retains: a survey he published in 2019 counted 31 unambiguous instances of parasitism frozen in Kachin amber against 4 from every compression deposit in the world combined. Feeding, mating and brood care are documented in the same way. Vaslin concedes two things. His counts come largely from museum holdings assembled by dealers who paid most for spectacular pieces, so the behavioural record is filtered a second time by the market. And the molecular promise raised in the early 1990s has evaporated: reports of ancient DNA recovered from insects in amber could not be repeated by other workers, and the polymer that protects a wing hair does nothing to hold a nucleic acid together.

EThe argument left the seminar room some time ago. Since 2020 several journals have obliged authors to document the provenance of Kachin material, because much of it has been mined and traded through an area at war, and papers unable to supply that documentation have been refused or withdrawn. Vaslin points out, uncomfortably for his own position, that the behavioural specimens he relies on are the ones dealers moved fastest and documented worst, so the evidence that most favours his case is the evidence least likely to clear the new bar. Sørholt answers that an appetite for the spectacular created both problems at once, since the taste that strips a war zone of its finest pieces also fills the literature with unrepeatable single specimens. Both regret one habit above all: that the field still reports species described rather than habits sampled, a figure that flatters every party to it.

FA settlement of sorts is emerging in which the two accounts are assigned to different questions rather than ranked against each other. For anything about the form or the conduct of a small forest animal, amber remains unmatched and its bias is beside the point, since a wasp gripping its host is not made less real by the absence of beetles that lived in soil. For anything about abundance or community composition, amber must be corrected against shale before it can be used at all, and at a third of sites it cannot be corrected. The unresolved question is whether resin chemistry itself altered as conifer forests gave way to flowering ones, which would make the filter vary through time as well as between trees; no series of measurements yet spans enough of the interval to settle it. What both researchers reject is the assumption behind the earliest surveys, namely that the fossils easiest to see are the ones that most deserve counting.