- The paper shows that free-swimming Thiovulum sp. ST self-organizes into persistent, periodic descending plumes and rising cell clouds through coupled chemotaxis, gravitaxis, and gyrotaxis rather than thermal forcing.
- The authors combine dark-field-like imaging and particle tracking across a 75 × 50 × 1 mm chamber to resolve rapid individual-cell motion, surface-layer formation within about 5 minutes, visible plumes near 20 minutes, and mature patterns after 30–50 minutes.
- The findings reveal chamber-spanning plumes with vertically layered vortices that persist for up to four hours, while leaving plume wavelength, termination, and cell egress as key targets for quantitative active-matter studies.
Overview
This paper documents and analyzes bioconvection in a colony of Thiovulum sp. ST, a recently isolated enrichment culture of large, fast sulfur-oxidizing bacteria confined in a Hele-Shaw chamber. The work accompanies a 2025 APS Division of Fluid Dynamics Gallery of Fluid Motion video award. The central observation is that a dense, initially homogeneous bacterial suspension spontaneously organizes into spatially periodic descending plumes interspersed with low-density rising clouds within minutes of homogenization — convection-like dynamics generated entirely by collective swimming rather than an imposed thermal gradient (2608.16145).
Physical mechanisms
The pattern-forming dynamics arise from the interplay of three orienting mechanisms rooted in the organism's metabolism. First, chemotaxis: like most sulfur oxidizers, Thiovulum sp. ST stores intracellular sulfur granules and is oxygen-limited, so cells swim up oxygen gradients toward the air–liquid interface where O2​ diffuses into the medium. Second, gravitaxis: negative buoyancy combined with a center-of-mass offset from the center of buoyancy (due to dense sulfur stores) produces a torque aligning swimming with gravity. Third, gyrotaxis: shear-induced torques further shape the narrow plumes. The accumulation of cells near the surface creates a density inversion that plausibly triggers a Rayleigh–Taylor-type instability, seeding the downward flows that initiate plume formation.
The phenomenology proceeds through distinct stages: (i) formation of a dense surface layer roughly 5 minutes after homogenization, with spatially periodic density fluctuations; (ii) emergence of visible plumes at regular intervals around 20 minutes, with finger-like structures that form, diffuse, and merge over minutes; and (iii) maturation after 30–50 minutes into stable, spatially periodic plumes spanning the full chamber height, with low-density "clouds" of upward-swimming cells between them. These structures persisted for the full 4-hour duration of the longest trials, limited ultimately by the availability of dissolved H2​S and O2​ rather than by any intrinsic decay of the instability. A notable observation is a "veil" of cells swimming parallel to high-density region boundaries, aligned with the chemical gradient, before cells descend, disperse, and resume upward chemotactic swimming — closing the circulation loop.
A striking structural finding is that although mature plumes span the chamber, the associated bioconvection vortices do not: distinct vertical levels of vortices form instead, with predominantly horizontal cell motion at the boundaries between levels. The paper explicitly concedes that what dictates where a plume terminates and how cells exit it remains unresolved.
Imaging and tracking methodology
To resolve individual cells over a large field of view (75mm×50mm×1mm chamber), the authors built a dark-field-like setup: two angled light panels behind the chamber, a black backdrop, and a tube extending past the macro-zoom lens to block unscattered light. This exploits the organisms' exceptional size and speed — diameters of 11μm and local swimming speeds up to 1mms−1, among the largest and fastest known free-swimming bacteria — and their bright appearance under incident illumination due to sulfur granules. Because the cells appear phototactic, illumination was kept uniform.
Time-lapse video was captured at 1-second intervals for several hours. Processing involved dividing each frame by a ten-image median-stacked background (removing dust, scratches, lighting irregularities), applying a short-wavelength bandpass filter to isolate cell-sized features, nonlinear intensity stretching, and least-squares displacement tracking linking features across consecutive frames. The resulting long particle tracks reveal the multi-scale structure: small-scale vortices around plumes near the surface and large-scale vortices filling the remainder of the chamber depth.
Relation to prior work
Unlike earlier experiments on tethered Thiovulum majus, where mucus-tethered cells collectively pumped nutrient-laden water through attached communities (2608.16145), here free-swimming Thiovulum sp. ST cells are observed moving parallel to strong oxygen gradients while forming patterns — behavior that may optimize simultaneous access to oxygen-rich and sulfide-rich water. Prior collective flows in similar bacteria have been shown to enhance oxygen transport efficiency, and the present system extends the catalog of bioconvection pattern-formers. The combination of large cell size, high contrast, rapid dynamics, and visible-to-the-unaided-eye collective structures makes this strain a practical model organism for active-matter studies analogous to flocking and schooling.
Limitations and open questions
The paper is primarily observational and descriptive. Several quantitative gaps remain open:
- Pattern wavelength control: the characteristic plume wavelength appears condition-dependent, but no systematic parameter study is presented relating wavelength to gravitactic torque, chemotactic sensitivity, gyrotaxis, or cell density.
- Plume termination and cell egress: the mechanism determining plume length and how cells leave a descending plume is unidentified, despite the observation of discrete vortex levels.
- Phototaxis: the reported light response is noted but uncharacterized, and uniform illumination is used as a mitigation rather than quantified.
- Data availability: supporting data are available only on request, limiting independent verification.
- The density-inversion/Rayleigh–Taylor interpretation of plume initiation is described as likely but not directly tested.
Quantitative measurements of pattern statistics under controlled variation of these parameters would provide direct tests for continuum active-matter theories of bioconvection.
Conclusion
The paper demonstrates that Thiovulum sp. ST sustains robust, self-organized bioconvection patterns in a quasi-two-dimensional geometry, driven by coupled chemotaxis, gravitaxis, and gyrotaxis rather than thermal forcing. High-contrast imaging plus computational tracking yields single-cell resolution across macroscopic scales, revealing layered vortex structures coexisting with chamber-spanning plumes. The principal open problems are predictive laws for plume wavelength, termination height, and vortex stratification as functions of taxis parameters and density.