The vast unknown microbial biosphere.

The desire to know how many living beings there are on the planet has been a permanent obsession of biologists. Many inventories have been done, especially with large animals and plants, since naturalists began modern taxonomy in the 18th century. But there is the feeling that we are missing a large part of the diversity of microorganisms. How large this ignorance is, however, is a matter of debate. In PNAS, Locey and Lennon (1) provide one such estimate that, if true, will have microbiologists struggling with the task of describing this diversity for decades, maybe centuries. Fig. 1A shows the distribution of primary producers on the planet, both on land and in the oceans. Despite the known difficulties of determining species and avoiding synonyms, we have a reasonable knowledge of where the ∼300,000 species of the main primary producers on land, the vascular plants, are (Fig. 1B). When we turn to the main primary producers in the oceans, however, we are at a loss to say how many species are there and how they are distributed. We are left with just the chlorophyll concentrations in Fig. 1A, because the marine primary producers are microbes. The contrast between our knowledge of the diversity of macrobes and microbes is thus apparent; the paper by Locey and Lennon (1) tries to put together data on both to estimate the total number of microbial species on Earth. Microbes and macrobes have seldom been considered together in the same study and with the same tools. Basically, considerations of the total number of species have ignored bacteria and archaea and estimates about protists have not been detailed enough. On the other hand, several extrapolation methods have been feeding a lively debate about the total number of species of insects and, particularly, of beetles. Erwin (2) created quite a stir with his estimate of 30 × 10 species of insects in tropical forests. This number was based on canopy fogging of 19 trees of Luehea seemannii. The tree canopy was packed in a bag and an insecticide fog spread inside. Then the dead insects were collected at the bottom. This was a major effort in itself, but it was just the beginning: 1,200 species of beetles were sorted out. Eventually, 162 species were considered to be specific to the particular tree host species. Next, a series of assumptions were needed to estimate the total number of insects. The number of tree species (50,000) was multiplied by this number of tree-specific beetles to give 8 × 10 species of canopy beetles. The proportion of beetles among insects was further used to estimate 20 × 10 arthropods in the canopy. Finally, the ground species were added to give a grand total of 30 × 10. This number was scary. Only about 1 × 10 species of insects had been described thus far; if the total were 30 × 10 the task of describing them seemed impossible. Since then, several other exercises have been carried out and the more optimistic scientists estimate that there may be “only” between 5 ± 3 × 10 (3) and 8.7± 1.3 × 10 species (4). These lower estimates have been used to argue that describing all of the species is possible in the next century in the face of increasing extinction rates (3). Of course the microbes were completely absent from such estimates. However, microbes rule the world. They are the most numerous, the most diverse, and the most important species in terms of carbon and nutrient cycling globally. How can a biologist speculate on the number of species on Earth and avoid considering microbes? The difficulty is that describing a microbial species is a tricky issue. In the case of bacteria and archaea, one needs to isolate it in pure culture, describe it biochemically, morphologically, and genetically, the 16S rRNA has to be sequenced, and the description has to be published in one of the few official bacterial taxonomy journals. This process is not that different from the effort necessary to describe a beetle. However, the main problem is that many bacterial and archaeal species are not easily isolated in pure culture. Some estimate than only 1% of the species present in an ecosystem can be retrieved in culture with current techniques. One shortcut came from sequencing of the 16S rRNA from natural communities. All of a sudden, this approach provided a wealth of novel microbial diversity that had remained beyond the reach of