AbstractThe rate of biomass growth throughout the cell cycle of prokaryotes is important in the study of global regulation. Two limiting cases have generally been considered: the exponential model and the linear model. The exponential model is a logical expectation because protein, the main component of biomass of a bacterial cell, increases continuously during the cell cycle and therefore the means for synthesis of other cell components and metabolites also increases. In addition, during the cell cycle, ribosomes, the means of production of proteins, increase monotonically. As a consequence, the increase of all should be autocatalytic and the content of cell substance should be an exponential function of time. Two cellular components would not be expected to increase exponentially: the DNA and the cell envelope. The former because of the intermittent synthesis of the chromosome, and the latter because of changes in the surface-to-volume ratio with growth and division. In contrast to the exponential model, the linear model of Kubitschek postulates that the cell only increases its membrane transport capability over a brief period during the cell cycle, and, thus limited by transport, all cell components can increase only at a constant linear rate during most of the cell cycle. Other proposed models are intermediate and assume that the growth rate of the cell depends on some cell cycle event, such as the initiation of chromosome replication.There are four classes of experiments that have been used to measure the accumulation of dry biomass or its components during the cell cycle of a bacterium, as typified byEscherichia coli. For the first class of experiments, the dimensions of living cells are measured under the microscope. So far, the experiments have been limited by the resolving power of the phase microscope, but adequate resolution should be possible with the confocal scanning light microscope or various video computer systems. Such experiments are calledintegralbecause augmentation of cell constituents is followed. The second class involves pulse-chase labeling of cells and then their separation into different phases of the cycle or age groups and measurement of the radioactivity per cell in the fractions. Such experiments are calleddifferentialin that the rate is measured directly instead of being deduced by comparing the total size at different times. A twofold difference is expected between these two limiting models for differential experiments instead of only 6% for the integral type of experiment. There are two major difficulties with published experiments of the differential class: one is the poor resolving power of all methods used to separate and analyze the cells in different phases of the cell cycle; the other is that prototrophic cells have been used without appropriate control experiments measuring the speed of feedback inhibition, completeness of inhibition, pool expansion, exchange with the medium, etc. All of these factors would be expected to vary with the surface-to-volume ratio and the number of uptake sites for the tracer, which must vary throughout the cell cycle. These problems could be overcome with suitable controls or the utilization of appropriate mutants blocked in the synthesis of the probe compound.The models have relevance to prokaryotes undergoing balanced growth; they may not be relevant to eukaryotic microbes or to eukaryotic cells in tissue culture that have endogenous rhythms or are controlled by protein growth factors. Logically, the models could possibly apply to a free-living cell that does not respond to environmental cues. Even under rigidly constant conditions, however, cells may try to respond to a stimulus that was periodic or regulatory under natural conditions, but is present at a constant level under the experimental culture condition.The third method analyzes the frequencies of cell size in growingbalancedpopulations and the distribution of dividing and newborn cells. The Collins-Richmond equation gives the mean growth rate of cells of various sizes. It appears to be valid in the central two thirds of the cell sizes in the population, but not at the extreme ends because small numbers of cells are usually enumerated and because a few aberrant (pathological) cells are commonly found. Another difficulty is the insufficient knowledge of the size distribution of dividing and newborn cells. The growth rate of various size cells also can be studied by the synthetic approach of Koch and Schaechter.With this method, it has been shown that the third mioment of the population size distribution (i.e., the skewness) is more determinative of the underlying growth law than are the higher and lower miments. A fouth method has been reported in preliminary form; this method measures the insertion of permease units into the membrane as a function of cell size. This fourth class of experiments should have an all-or-none difference between the exponential model and the linear model.The bulk of this review is critical of all phblished work in the field. However. the conclusion of this review is that adequate experimental approaches exist and, with feasible refinement and generalization, experiments of any of the four classes could answer definitively the quesion as to how a bacterium grows. Within limitations, the availabel data exclude the linear model and support models that approximate exponential growth of cell biomass within the cell cycle.