Mathematics in Science: Physics, Chemistry, and Biology Connections

Mathematics sits inside every scientific measurement, model, and prediction — not as decoration, but as the load-bearing structure. This page examines how mathematical frameworks connect to physics, chemistry, and biology, what specific tools each discipline borrows, and where those connections are most consequential. The relationships are not symmetrical: physics runs deepest on calculus and differential equations, chemistry leans on probability and statistical mechanics, and biology has only recently committed fully to quantitative methods.

Definition and scope

When scientists say a theory is mathematically rigorous, they mean something precise: the theory's predictions are derived from equations whose terms have defined units, known ranges of validity, and testable numerical outputs. The National Science Foundation frames quantitative reasoning as a cross-cutting competency across all funded STEM disciplines, which signals how central mathematics has become to scientific legitimacy itself.

The scope here covers three distinct but overlapping relationships. First, mathematics as a language — the notation and logical structure that lets scientists communicate precisely. Second, mathematics as a modeling tool — using equations to represent physical, chemical, or biological systems. Third, mathematics as a discovery engine — cases where following the math led to phenomena observed only afterward. Paul Dirac's prediction of antimatter from the mathematics of quantum mechanics, before positrons were detected experimentally, is the canonical example of the third category.

These relationships span the full breadth of applied mathematics, from the deterministic equations of classical mechanics to the stochastic models now standard in population genetics.

How it works

The mechanism differs by discipline, but a common architecture runs through all three sciences:

  1. Identify a system — isolate the phenomenon worth modeling (a falling object, a reacting molecule, a spreading infection).
  2. Assign variables — attach mathematical symbols to measurable quantities, with explicit units.
  3. Write governing equations — express the relationships between variables, often as differential equations.
  4. Solve or simulate — find analytical solutions where possible; use numerical methods where not.
  5. Compare to data — measure how well predictions match observation, using statistics and probability to quantify the fit.
  6. Revise the model — update assumptions when the residuals are too large.

Physics is the most mathematically intensive of the three. Newton's second law (F = ma) is a second-order differential equation the moment acceleration is expressed as the derivative of velocity. James Clerk Maxwell's four equations of electromagnetism, published in 1865, unified electricity, magnetism, and light into a single mathematical framework — and directly predicted that electromagnetic waves travel at approximately 3 × 10⁸ meters per second, a figure that matched measured light speed.

Chemistry reaches for mathematics most visibly in quantum chemistry and thermodynamics. The Schrödinger equation governs electron behavior in atoms and molecules; solving it exactly is possible only for hydrogen. For every other element, chemists use approximation methods — perturbation theory, density functional theory — that are themselves deep mathematical structures. The Boltzmann distribution, which describes how molecules distribute energy at a given temperature, is an exponential function derived from statistical mechanics and published by Ludwig Boltzmann in the 1870s.

Biology arrived later. The Hardy-Weinberg principle (1908), which describes allele frequency equilibrium in a population, was one of the earliest algebraic models in genetics. The field accelerated when Alan Turing published his 1952 paper on reaction-diffusion equations to explain biological pattern formation — the mathematics behind why a zebra has stripes and a leopard has spots. Computational biology now uses linear algebra and graph theory to map protein interactions across networks containing thousands of nodes.

Common scenarios

Three scenarios illustrate how this plays out in practice:

Pharmacokinetics in medicine. When a drug enters the body, its concentration over time follows an exponential decay described by first-order differential equations. Clinicians use these models to set dosing intervals — the half-life of ibuprofen in adults is approximately 2 hours, which is a mathematical constant derived from measured clearance rates, not a clinical judgment call.

Climate modeling. The equations governing atmospheric circulation are partial differential equations discretized onto a grid. The Intergovernmental Panel on Climate Change (IPCC) bases its projections on ensembles of such models. The uncertainty ranges in temperature projections — typically expressed as a confidence interval around a central estimate — are statistical outputs from those runs.

Genomic sequencing analysis. Aligning a sequenced genome against a reference genome is a combinatorial optimization problem. Tools like BLAST (Basic Local Alignment Search Tool), developed by the National Center for Biotechnology Information, use scoring matrices derived from probability theory to find the best-fit alignment across billions of base pairs.

Decision boundaries

Not every scientific question is equally amenable to mathematical treatment, and understanding the boundaries matters as much as knowing the tools.

Deterministic vs. stochastic models. Classical mechanics is deterministic: given initial conditions, the outcome is fixed. Quantum mechanics is irreducibly probabilistic. Population ecology sits between the two — predator-prey models like Lotka-Volterra are deterministic differential equations, but real populations are subject to random fluctuations that require stochastic extensions.

Analytical vs. numerical solutions. When equations can be solved in closed form, the result is exact and portable. When they cannot — which is most of the time in chemistry and biology — numerical simulation introduces approximation error that must be tracked. This distinction shapes the entire field of mathematical modeling.

Continuous vs. discrete frameworks. Physics often treats matter as continuous; biology must reckon with discrete individuals, discrete generations, and discrete genetic variants. Discrete mathematics and combinatorics become essential the moment the model counts organisms rather than measuring concentrations.

The entire foundation of the mathematics underlying scientific work rests on these distinctions — choosing the wrong framework produces predictions that are precise but wrong, which is arguably worse than no prediction at all.

References