Gas Bubbles in Fossil Amber as Possible Indicators of the Major Gas Composition of Ancient Air

AMBER IS ANCIENT TREE RESIN THAT has hardened and been preserved in sedimentary rocks; its age ranges from Carboniferous ( 300 million years ago) to Pleistocene-Recent (1). Besides bearing paleobiological information in its chemical structure and in its embalmed insects and spores (1), amber commonly contains gas bubble inclusions. From preliminary data it has been suggested (2-4) that these bubbles may represent mainly ancient air trapped at the time the original resin issued forth from its host tree. We present data that support this hypothesis and suggest that the major gas composition of air (N2/O2 ratio) has changed appreciably during the past 90 million years.

Amber samples studied include: (i) Dominican amber (from the Dominican Republic) of probable late Oligocene-to-Early Miocene age (5), (ii) Baltic amber of probable Oligocene-Eocene age (1, 6), and (iii) Upper Cretaceous amber from Cedar Lake, Manitoba, whose age (6, 7) is estimated to be "Senonian" (75 to 95 million years ago). Hardened modern pine tree resin from New Zealand (Agathis australis) was also studied as a check on our method. The Dominican amber consisted of centimeter-sized chunks containing millimeter-sized bubbles (Dominican 1 and 3), and a 4- to 5-cm diameter, transparent, gem-quality sample (Dominican 2) with a few large millimeter-sized bubbles trapped deeply within the clear amber (8). The Baltic material and New Zealand resin consisted of centimeter-sized chunks with bubbles readily visible under a binocular microscope. The Upper Cretaceous amber consisted of small pieces of clear amber, 2 to 4 mm in size, which in thin section were found to contain small bubbles ranging in size from about 10 to 100 m.

Analysis of the amber was done by crushing it under vacuum and analyzing the released gases by time-resolved quadrupole mass spectrometry. The detection limit was about 10 ppm and the accuracy was about 3 to 5% of the amount of each gas present (4, 9). We used two crushing methods. The first method, applied to Dominican samples 1 and 2 only, consisted of placing bubble-rich chunks (with diameters of 1.5 to 2 cm) in stainless steel tubes that were attached to the mass spectrometer and evacuated to less than 10(-6) torr. The tubes were sealed, removed, and the contained amber broken to mainly sand-sized fragments by gently crushing the tube in a hydraulic press. The released gases in the still-evacuated tubes were then transferred to the mass spectrometer for analysis. Because of the large bubble size, this crushing technique resulted in the release of much larger quantities of gas ( 10(-7) mole/g) as compared to the second method. The second method, applied to all other samples, involved successive, but very gentle, crushes in a small, evacuated (10(-9) torr initial pressure), piston-cylinder crushing chamber that was attached at all times to the mass spectrometer. By slowly turning a crank, the degree of breaking and gas release could be controlled and small amounts of gas, on the order...