Identifying bacteria on agar plates is a layered process that starts with your eyes and progressively adds laboratory tests until you narrow down the organism. No single observation gives you a definitive answer. Instead, you build a profile: colony shape, size, color, texture, how the organism behaves on specialized media, how it reacts to staining, and what it does in a handful of quick biochemical tests. Together, these clues function like a decision tree, each branch eliminating possibilities until you arrive at a genus or species. The process has been refined over more than a century, and while modern instruments can now confirm an identification in minutes, the fundamentals of reading a plate remain surprisingly hands-on.
Reading Colony Morphology
The first step is simply looking at what grew. When bacteria multiply on the surface of agar, they form visible clusters called colonies, and the physical characteristics of those colonies vary enough between species to be genuinely useful. The American Society for Microbiology recommends evaluating colonies based on size, color, texture, edge shape, and elevation as a standard protocol for distinguishing phenotypically different bacteria.1MDPI. Distinction of Different Colony Types by a Smart-Data-Driven Tool A round, smooth, glistening colony that is a couple of millimeters across tells you something very different from a flat, irregular, dry-looking colony that spreads across the plate.
Here are the main features to evaluate when you pick up a plate:
- Size: Measure or estimate in millimeters. Some species produce pinpoint colonies even after overnight growth, while others spread aggressively.
- Shape: Circular, irregular, filamentous, or rhizoid (root-like). Most common clinical isolates form roughly circular colonies, so anything that deviates is a useful clue.
- Color: White, cream, yellow, gold, green, or even red. Many bacteria produce pigments that make their colonies distinctive.
- Surface texture: Smooth (shiny), rough (matte or granular), mucoid (wet and glossy, suggesting a heavy capsule).
- Edge: Entire (clean border), undulate (wavy), lobate (lobed), or filamentous (hair-like projections).
- Elevation: Flat, raised, convex, or umbonate (raised with a central bump).
None of these features alone pins down a species, but together they can strongly suggest one. A golden-yellow, convex, smooth colony on nutrient agar immediately raises suspicion for Staphylococcus aureus, for example, while a large, flat, spreading colony with an irregular edge might point toward a Bacillus species.
What Hemolysis Tells You on Blood Agar
Blood agar plates contain red blood cells mixed into the medium, and the way bacteria break down those cells produces visible patterns around each colony. This hemolytic reaction is one of the most informative quick observations in clinical microbiology. Three patterns matter:
- Beta-hemolysis: A clear, transparent zone around the colony where red blood cells have been completely destroyed. Classic for group A streptococci and Staphylococcus aureus.
- Alpha-hemolysis: A greenish or brownish discoloration around the colony, indicating partial breakdown of hemoglobin. Typical of Streptococcus pneumoniae and viridans streptococci.
- Gamma-hemolysis: No change at all. The organism does not lyse red blood cells. Enterococci and some other species fall into this category under certain conditions.
Incubation conditions can shift hemolytic patterns, which is worth knowing. Enterococcus species, for instance, consistently produce alpha-hemolysis when incubated in a carbon dioxide atmosphere, but their hemolytic reaction becomes variable under anaerobic conditions. Beta-hemolytic streptococci, by contrast, produce the same clear hemolysis regardless of the atmosphere.2PubMed. Pig and goat blood as substitutes for sheep blood in blood-supplemented agar media This means the same organism can look different on a blood plate depending on how it was incubated, so experienced technologists pay attention to the conditions, not just the zone.
Selective and Differential Media
Plain nutrient agar grows almost anything, which is sometimes the point but often not. In clinical and food microbiology labs, plates are chosen specifically to suppress certain organisms while highlighting others. These specialized formulations fall into two overlapping categories: selective media (which inhibit growth of unwanted bacteria) and differential media (which make target organisms visually distinct from everything else on the plate). Many commonly used plates are both.
MacConkey agar is a workhorse example. It contains bile salts and crystal violet, which suppress most gram-positive bacteria, allowing gram-negative organisms to grow essentially unopposed. It also contains lactose and a pH indicator, so you can immediately see whether a colony ferments lactose. Lactose fermenters like E. coli produce acid that turns colonies pink or red, while non-fermenters like Salmonella stay pale or colorless.3Journal of Physics: Conference Series. Growth Analysis of Escherichia coli and Salmonella typhi on MacConkey Agar Modification That single color difference on one plate immediately sorts gram-negative isolates into two broad camps.
Chromogenic media take this concept further by incorporating enzyme-specific substrates that release color when a target organism cleaves them. The detection of a specific enzyme for E. coli enumeration is well established, and newer chromogenic plates can distinguish organisms like Salmonella, enterococci, Listeria monocytogenes, Staphylococcus aureus, and Clostridium perfringens based on colony color alone.4PubMed. New developments in chromogenic and fluorogenic culture media Some of these plates can give you a presumptive identification in 18 to 24 hours without any additional testing, which is a significant time savings in a busy lab.
Pigment Production as a Visual Shortcut
Some bacteria produce pigments so distinctive that their color is nearly diagnostic on its own. The most famous example is Pseudomonas aeruginosa, which produces pyocyanin, a blue-green pigment synthesized by roughly 90 to 95 percent of strains.5PubMed Central. Pseudomonas aeruginosa’s greenish-blue pigment pyocyanin: its production and biological activities Pyocyanin diffuses into the surrounding agar, so the medium itself takes on a greenish-blue tint around and beneath the colonies.6Scientific Reports. Isolation and characterization of nutrient dependent pyocyanin from Pseudomonas aeruginosa and its dye and agrochemical properties If you see that color spreading through the plate, Pseudomonas aeruginosa jumps to the top of your suspect list before you run a single test.
Other examples are less dramatic but still useful. Serratia marcescens can produce a red pigment called prodigiosin, though this is temperature-dependent and not always present. Staphylococcus aureus colonies often have a golden hue. Chromobacterium violaceum produces a vivid violet pigment. These colors are not confirmatory on their own, but they narrow the field considerably and tell you which follow-up tests to run.
Swarming and Motility Patterns
Most bacteria sit politely within their colonies. Proteus mirabilis does not. On standard agar, P. mirabilis swarms outward in waves, creating a distinctive bull’s-eye pattern of concentric rings that can cover an entire plate.7PubMed Central. Characterization of Proteus mirabilis precocious swarming mutants: identification of rsbA, encoding a regulator of swarming behavior Each ring represents a cycle where cells elongate, migrate outward as a group, then pause and consolidate before the next wave.8PubMed Central. Transcriptome of swarming Proteus mirabilis The result is unmistakable: when you open an incubator and find a plate that has been overtaken by concentric ripples, Proteus is almost certainly the culprit.
The degree of swarming depends partly on the agar’s rigidity. On lower-concentration hard agar, cell elongation and collective migration are tightly linked and the bull’s-eye pattern is pronounced. At higher agar concentrations, these behaviors become decoupled and swarming is more constrained.9PubMed Central. Cell Shape and Population Migration Are Distinct Steps of Proteus mirabilis Swarming That Are Decoupled on High-Percentage Agar Labs sometimes deliberately use higher-concentration agar to suppress swarming when they need to isolate individual colonies from a mixed specimen that may contain Proteus.
Gram Staining After You Pick a Colony
Once you have chosen a colony of interest based on its appearance, the next step is almost always a Gram stain. This technique, dating to 1884, divides bacteria into two large groups based on their cell-wall structure. A crystal violet–iodine complex is applied, then washed with alcohol. Bacteria with thick cell walls retain the purple dye and are classified as gram-positive. Those with thinner walls lose the purple dye during the alcohol wash, pick up a pink safranin counterstain instead, and are classified as gram-negative.10PubMed. Gram staining
Beyond the color split, the Gram stain also shows you cell shape and arrangement. You can see whether you are dealing with cocci (round cells), bacilli (rod-shaped cells), or something in between. Cocci arranged in grape-like clusters suggest staphylococci; cocci in chains suggest streptococci; gram-negative rods open up a wide differential that includes E. coli, Klebsiella, Salmonella, and dozens of others. The Gram stain does not identify a species, but it instantly halves the bacterial kingdom and tells you which set of follow-up tests to reach for.
Quick Biochemical Tests
After colony morphology and the Gram stain, a handful of rapid bench tests can push you much closer to a species-level identification. These tests detect specific enzymes or metabolic capabilities and typically give results within seconds to a few hours.
The catalase test is often the very first one performed on gram-positive cocci. You drop hydrogen peroxide onto a colony and watch for bubbles. Staphylococci produce catalase and bubble vigorously; streptococci and enterococci do not. Within the gram-negative rod family, catalase reactions vary in strength: Serratia, Proteus, and Providencia are vigorous catalase producers, while Escherichia and Shigella strains are mostly non-reactive or weakly reactive.11PubMed Central. Catalase test as an aid to the identification of Enterobacteriaceae
From there, the choice of follow-up tests depends on what the Gram stain showed. For gram-positive cocci in clusters that test catalase-positive, a coagulase test separates S. aureus (coagulase-positive) from other staphylococci. For gram-negative rods, tests like indole production, citrate utilization, urease activity, and the triple sugar iron (TSI) reaction help sort through the large family of gut-associated bacteria. A lactose-fermenting gram-negative rod that is catalase-positive, indole-positive, citrate-negative, and urease-negative, with an acid-over-acid TSI reaction producing gas but no hydrogen sulfide, fits the profile of E. coli.12Journal of Surgical Case Reports and Images. Overview on Old and New Biochemical Test for Bacterial Identification These flowchart-style decision trees have been used in teaching and clinical labs for decades, and they remain effective even in the era of automated instruments.
Smell as an Underrated Clue
Experienced microbiologists sometimes recognize organisms by their odor before they even look at colony morphology. Pseudomonas aeruginosa is often described as smelling like grape juice or corn tortillas. E. coli has a fecal odor that is hard to mistake once you have encountered it. Proteus species smell strongly of ammonia. Some anaerobes produce a foul, sulfurous stench. Odor-based identification is subjective and not standardized, but it is widely acknowledged in teaching labs. Research has specifically selected microbial strains with distinct odors for use in identification accuracy studies, recognizing smell as a real diagnostic observation.13PubMed Central. Accuracy of odor-based microorganism identification by microbiological technologists with different years of experience: A cross-sectional study
That said, smelling plates deliberately is discouraged in safety protocols. You should never hold a plate up to your nose and inhale. The standard practice, if you notice an odor at all, is to gently waft air from the plate toward you with your hand. And smell alone never substitutes for a proper identification workup. It is best thought of as a nudge: it points your attention toward the most likely candidate and helps you decide which tests to prioritize.
Why the Same Bacterium Can Look Different on Different Plates
One of the trickiest aspects of plate reading is that colony morphology is not fixed. The same strain of bacteria can produce colonies that look remarkably different depending on growth conditions. A study using two strains of Pseudomonas aeruginosa demonstrated that variables like colony growth time, plate density, the composition of the culture medium, and whether the bacteria came from a free-floating or biofilm culture all significantly changed the appearance of colonies. Across these conditions, 18 distinct morphotypes were identified from just two strains, differing in size, shape, color, texture, and margin.14PubMed. Improvements on colony morphology identification towards bacterial profiling Growth time and medium composition caused the largest shifts.
This variability is why experienced microbiologists treat colony morphology as a starting point, not an endpoint. A colony that looks “wrong” for a given species may simply have been incubated longer, grown on a different medium, or come from a crowded plate where nutrients were depleted. When colonies do not match what you expect, the answer is not to force-fit an identification. The answer is to subculture a pure colony onto fresh medium, re-examine it, and then run confirmatory tests.
Growing Anaerobes and Atmosphere-Dependent Organisms
Not all bacteria grow in the presence of oxygen. Anaerobic bacteria, which include clinically important organisms like Clostridium and Bacteroides, require oxygen-free environments. Labs achieve this using a range of systems: anaerobic pouches, sealed jars with gas-generating packets, anaerobic boxes, and full walk-in anaerobic chambers. These systems provide suitable conditions to successfully isolate even strict anaerobes from clinical specimens.15PubMed Central. Biosafety guidelines for handling microorganisms in the teaching laboratory: development and rationale If you plate a wound specimen on blood agar and incubate it only in a standard aerobic incubator, you will miss any anaerobes entirely, which can mean missing the actual cause of an infection.
Some organisms also require elevated carbon dioxide or microaerophilic conditions (reduced oxygen). Campylobacter, for example, grows best in about 5 percent oxygen and 10 percent COâ‚‚. Choosing the right atmosphere is as important as choosing the right medium, and when plates come back “no growth” on a specimen that clinically looks infected, the first question is often whether the right incubation conditions were used.
Instrument-Based Identification
For decades, the workflow above was essentially all there was: plate the specimen, look at the colonies, stain them, run biochemical tests, and build your identification from the accumulated results. Modern labs still do all of this, but they also have access to instruments that can confirm or replace portions of the process.
MALDI-TOF mass spectrometry has transformed clinical microbiology over the past 15 years. The concept is straightforward: you smear a tiny amount of a colony onto a target plate, hit it with a laser, and the instrument measures the mass of the proteins released. The resulting protein fingerprint is compared against a database. Applying bacterial colonies directly to the instrument without additional extraction steps produces rich, reproducible peak patterns that allow highly accurate identification.16Molecular & Cellular Proteomics. Highly Efficient Classification and Identification of Human Pathogenic Bacteria by MALDI-TOF MS The technique is reliable enough to identify difficult organisms like Brucella to the genus level from both culture plates and directly from blood culture bottles.17PLoS ONE. Identification of Brucella by MALDI-TOF Mass Spectrometry. Fast and Reliable Identification from Agar Plates and Blood Cultures Results come back in minutes rather than hours or days.
When MALDI-TOF cannot resolve an identification or when phenotypic methods give conflicting results, molecular techniques step in. Sequencing the 16S ribosomal RNA gene has become a standard tool for resolving ambiguous cases. One study found that conventional biochemical methods had misidentified clinical isolates, and 16S gene sequencing corrected those errors, identifying the organisms as Enterococcus faecalis, Cardiobacterium valvarum, and Streptococcus mutans respectively. The researchers concluded that the original misidentifications could have misled clinicians and affected patient care.18PubMed Central. The role of 16S rRNA gene sequencing in identification of microorganisms misidentified by conventional methods Gene sequencing is unaffected by the morphological variability and phenotypic oddities that trip up conventional methods.
Automated Image Analysis and Machine Learning
A newer frontier involves using cameras and algorithms to do what human eyes have done for over a century: look at colonies and classify them. Machine learning approaches have been applied to a range of bacterial image types, including whole agar plate photographs, hyperspectral images, and microscopic images of stained specimens.19PubMed Central. Automated Bacterial Classifications Using Machine Learning Based Computational Techniques: Architectures, Challenges and Open Research Issues These systems learn the visual features that distinguish one type of colony from another and can flag colonies of interest or suggest identifications.
Recent work has shown impressive accuracy. A deep learning model tested on high-resolution plate images achieved detection accuracy ranging from 0.944 to 0.996 across 19 bacterial species, with minimal variation between categories.20Scientific Data. High-Resolution Colony Images of Clinically Isolated Bacteria for Automated Detection and Deep Learning Earlier studies using convolutional neural networks for colony classification also demonstrated that deep learning can extract texture and morphological features that are difficult for humans to articulate but consistently distinguish between genera and species.21PLOS ONE. Deep learning approach to bacterial colony classification These tools are not replacing microbiologists, but they are beginning to handle the screening step, particularly in high-volume labs where thousands of plates pass through each day.
Safety When Handling Plates
Whenever you are working with bacteria on agar plates, even in an educational setting, biosafety practices matter. Guidelines adapted from the CDC’s Biosafety in Microbiological and Biomedical Laboratories provide best practices for handling microbes at biosafety levels 1 and 2.15PubMed Central. Biosafety guidelines for handling microorganisms in the teaching laboratory: development and rationale At a minimum, this means working near a flame or in a biosafety cabinet, wearing gloves, never eating or drinking in the lab, and properly disposing of used plates by autoclaving.
For plates that may contain unknown organisms from clinical or environmental samples, treat them as potentially hazardous until identified. Unknown gram-negative rods from a clinical specimen could turn out to be anything from harmless E. coli to something requiring more stringent containment. The identification process itself shapes the safety response: once you know what you are dealing with, you can adjust your precautions accordingly. Until then, default to the higher level of caution.
How Colony Shape Changes in Three Dimensions
Most identification guides describe colonies as if they are two-dimensional objects viewed from above, but colonies are three-dimensional structures, and their growth in the vertical dimension follows interesting patterns. Research on bacterial colonies growing in soft gel matrices has shown that expanding colonies undergo a morphological instability where the colony surface roughens over time, eventually adopting a highly branched, broccoli-like interface. This instability appears within just a few hours and has been observed across diverse species including E. coli, Vibrio cholerae, Pseudomonas aeruginosa, and Komagataeibacter sucrofermentans, suggesting it is a general physical phenomenon rather than something unique to a particular organism.22Proceedings of the National Academy of Sciences (PNAS). Morphological instability and roughening of growing 3D bacterial colonies
For practical identification purposes, this means that older or very thick colonies may develop surface textures that have more to do with physics than with biology. A colony that looked smooth at 18 hours might appear rough or irregular at 48 hours, not because the bacterium changed but because the colony outgrew the conditions under which it could maintain a smooth surface. Recognizing this helps avoid the trap of assuming a rough older colony is a different organism than the smooth younger one growing next to it.