Gravitational Lensing Tests

Gravitational lensing tests how spacetime guides light and coherent waves. Its distinctive strength is sensitivity to the sum of the two weak-field metric potentials; dynamics primarily probes the time potential. Its distinctive danger is projection: every image, shear map, delay, or interference pattern also depends on source structure, lens mass, line-of-sight matter, selection, and calibration. This chapter develops the observables and likelihoods needed to preserve the comparison. Its empirical status is current through 2026-08-19.

Helpful background. Geodesics and Jacobi Fields supplies ray and bundle propagation; Probability and Statistical Inference supplies normalized likelihoods and selection; Weak-Field Gravity supplies the two-potential metric; and Relativistic Cosmology supplies distance, growth, and light-cone conventions.

Enter the lensing comparison

In conformal Newtonian gauge,

$$ds^2=-(1+2\Phi)dt^2+a^2(1-2\Psi)d\boldsymbol{x}^2.$$

Slow motion responds to $\Phi$; leading lensing responds to $\Phi+\Psi$. This complementarity motivates gravitational-slip tests, but neither field is observed directly. A weak-lensing catalog contains calibrated galaxy shapes. A strong lens supplies pixels, delays, spectra, and environment data. A microlensing survey supplies selected light curves. A gravitational-wave detector supplies calibrated strain. Each requires a forward map from gravity and sources through instrument response and selection into data.

Across the chapter, a valid conclusion follows the same chain:

1. declare the metric or propagation deformation and its scale and environment;
2. predict a lens map, shear statistic, arrival time, magnification, or complex amplification;
3. specify source, lens, line-of-sight, and population variables;
4. pass the prediction through response, cadence, masks, thresholds, and quality cuts;
5. fit a normalized likelihood with priors and covariance; and
6. report only the gravity combination that survives adversarial source and selection alternatives.

A reconstructed mass map is not a direct measurement of the metric. Likewise, a visually striking arc, a long transient, or a waveform modulation is not a gravity test until the corresponding unlensed, astrophysical, and instrumental hypotheses have been compared.

For a first reading, the following small set of symbols connects the chapter. Each leaf then supplies the derivation and its assumptions.

Quantity	Meaning	What it must not be confused with
$\widehat\alpha$	Physical bending of a ray near one lens	Reduced image–source separation $\alpha=(D_{ls}/D_s)\widehat\alpha$
$\kappa,\gamma$	Local convergence and shear of an image	A direct three-dimensional mass measurement
$\mu$	Signed geometric image magnification	Wave amplitude, which scales as $\sqrt{\lvert\mu\rvert}$
$\Delta t$	Differential arrival time between images	A lens distance without a Fermat-potential model
$\theta_E,t_E$	Einstein angle and angular crossing time	Lens mass inferred from duration alone
$F(f)$	Complex wave amplification	A frequency-independent flux magnification

Start with the two-potential and thin-lens geometry, then choose shear statistics for surveys, mass-sheet transformations for resolved strong lenses, or the point-lens scale calculation for microlensing. Read gravitational-wave association before interference: one first needs to know whether signals share a source, and only then what their propagation can test.

Route through six scales

The pages proceed from the two-potential foundation to increasingly specialized data:

1. Lensing and Dynamical Potentials derives why light measures $\Phi+\Psi$ while slow tracers measure $\Phi$, and states the gauge, quasistatic, projection, and screening limits of a slip parameter.
2. Weak-Lensing Gravity Statistics builds tomographic shear and cross-correlation likelihoods with shear response, redshift, intrinsic alignment, baryons, galaxy bias, masks, and covariance.
3. Strong-Lensing Mass and Time-Delay Tests develops the Fermat-potential observable and shows why mass-sheet, source-position, anisotropy, and external-convergence freedom require imaging, delays, kinematics, and environment data together.
4. Microlensing and Compact Lenses separates duration from mass and explains how parallax, finite-source structure, lens flux, astrometry, Galactic priors, and detection efficiency change compact-object membership.
5. Gravitational-Wave Lensing Tests distinguishes common-source association from a propagation comparison and pins the difference between peer-reviewed O3 results, a submitted O4a search, and current GWTC-5.0 data inputs.
6. Wave-Optics and Interference Tests derives the diffraction integral, its geometric limit, and the coupled amplitude–phase checks needed to distinguish a lens from waveform error or noise.

The first page is analytic and durable. The remaining pages contain dated empirical boundaries because catalogs, calibration, population models, and null searches change. A later dataset should not be inserted into one conclusion while leaving its response, selection, covariance, and software products behind.

Shared degeneracies and decisive controls

Three degeneracies recur. First, lensing measures a projected Weyl potential while dynamics reconstructs a three-dimensional time potential; triaxiality and disequilibrium can imitate slip. Second, source transformations preserve much of the observed lens map: the mass-sheet transformation rescales delays, while blending or unknown proper motion mixes microlens mass with population. Third, catalog construction depends on lensing itself through magnification, cadence, template mismatch, and threshold crossing.

The corresponding controls are joint and predictive. Weak-lensing pipelines should recover blinded parameters through image simulation and pass B modes and catalog splits. Strong-lens models should recover synthetic mass-sheet transformations after adding resolved kinematics and environment information. Microlensing analyses should inject events through the detection pipeline and demonstrate population coverage under alternate Galactic priors. Gravitational-wave searches should calibrate common-source odds using both lensed and unrelated injections and require consistent sky, delay, phase, and lens predictions.

These controls do not guarantee a unique theory. They bound the conclusion to a parameterization, source family, scale, redshift, and environment. Screening may make a dense lens obey its GR limit while cosmological growth differs. A null wave-optics search may miss lenses whose delay scale lies outside the coherent detector band. The strongest evidence is therefore a common parameter map that predicts several observables under shared nuisance parameters, rather than a small residual in one summary statistic.

What current results establish

The mature foundation is the lens and wave-propagation formalism: null bundles, Fermat delays, the mass-sheet transformation, and diffraction limits. Survey and catalog results are conditional. DES Year 3 provides a documented weak-lensing data and calibration example; TDCOSMO shows how lens-population flexibility changes time-delay inference; OGLE demonstrates the distinction between public alerts and a selection-complete population sample; and LVK’s peer-reviewed full-O3 lensing search reports no compelling evidence within its tested models. GWTC-5.0 is the current event-data boundary at the cutoff, not automatically a new lensing result.

Consequently, the chapter does not claim a source-independent gravitational anomaly. It explains how one could become credible: the signal must survive alternate source and lens families, end-to-end selection, held-out observables, and an independently constrained environment. Mass-sheet freedom, line-of-sight structure, baryonic modeling, microlens populations, and false pair association remain leading limitations.

Review the chapter

1. Potential comparison. A lensing analysis and a stellar-dynamical analysis report different enclosed masses. Why is their ratio not automatically a measurement of gravitational slip?

Solution

Lensing reconstructs a projected combination of $\Phi+\Psi$, whereas stellar dynamics infers the three-dimensional $\Phi$ through a tracer and equilibrium model. Distance, geometry, line-of-sight convergence, stellar anisotropy, mass-to-light ratio, and selection can shift their inferred masses differently. Therefore the two datasets must be predicted from one latent source with shared covariance. Only the residual combination, stable across these alternatives and within a declared scale and screening regime, constrains slip.

2. Catalog reasoning. Why does a current gravitational-wave catalog not by itself update a gravitational-wave lensing conclusion?

Solution

A catalog supplies selected candidates, strain, and parameter estimates, but a lensing conclusion additionally requires a pair or distortion search, calibrated prior odds, detector duty cycle, subthreshold efficiency, lens and source populations, waveform and noise alternatives, and injection controls. Because those elements are analysis-specific, adding events without rerunning the complete selection-normalized likelihood cannot preserve the old interpretation. A new catalog provides inputs for repeating the lensing analysis; it does not by itself establish a lensing result.

Further reading

• Bartelmann, M., and Schneider, P. (2001), “Weak Gravitational Lensing,” Physics Reports 340, 291–472. DOI.
• Birrer, S., et al. (2020), “TDCOSMO IV: Hierarchical Time-Delay Cosmography—Joint Inference of the Hubble Constant and Galaxy Density Profiles,” Astronomy & Astrophysics 643, A165. DOI.
• LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration (2024), “Search for Gravitational-Lensing Signatures in the Full Third Observing Run of the LIGO–Virgo Network,” Astrophysical Journal 970, 191. DOI.
• Takahashi, R., and Nakamura, T. (2003), “Wave Effects in the Gravitational Lensing of Gravitational Waves from Chirping Binaries,” Astrophysical Journal 595, 1039–1051. DOI.