Homogeneity rules fit products created by mixing a batch and dividing it into servings; they fit hand-assembled composites only through sampling conventions. An edible mass can be blended until a collected sample represents the whole. A moon rock is built as a set of separate objects from flower, concentrate and kief. Composite testing can produce a valid batch average, yet the rule's original physical model—a mixed volume with uniform distribution—does not describe the way those pieces are manufactured.

Massachusetts makes the mixed-product model visible in its testing guidance. The Cannabis Control Commission's Administrative Order No. 4 materials address production batches, sample collection and final-form testing. Later protocol recommendations instruct laboratories to homogenize a received test sample before taking analytical subsamples. Homogenization at the laboratory makes the submitted material suitable for measurement. It cannot retroactively mix finished units across the processor's entire batch.

Oregon's sampling guide distinguishes harvest lots from process lots and describes how finished cannabinoid products move through compliance testing. A process lot groups material of the same type made at the same time with the same methods, procedures and ingredients. That definition works cleanly for mixed liquids and edible masses. For assembled pieces, sameness resides in method and input set while the actual ratio on each unit emerges from separate handling events.

Colorado's production-batch framework adds the final-form requirement. Regulated products are grouped and sampled under rules that define a production batch and the material sent to a testing facility. The laboratory reports the submitted sample according to validated methods. The compliance result supports release of the batch. It does not create a unit-by-unit map, because the laboratory receives only the selected sample and then prepares it for analysis.

Composite sampling is the practical bridge. Several finished pieces can be selected according to the sampling plan and combined before analysis. The result describes the combined material faithfully. Its strength is coverage across more than one object; its limitation is that averaging erases the distance between them. Two sets of pieces can produce the same average while carrying different spreads in component ratio. The compliance number and the manufacturing distribution answer different questions.

Process control can supply the missing view without changing the laboratory rule. A producer can measure component pickup, finished unit weight, coating coverage and other physical tolerances at intervals through a run. Those checks remain manufacturing records rather than consumer claims. They show whether assembly stayed centered and how far individual observations moved from the target. A well-defined batch then has two layers of evidence: a compliance result from the laboratory and a distribution record from the bench.

The distinction also clarifies product classification. A composite piece is neither loose flower nor a mixed concentrate. It contains both in a stable assembled form. Applying a rule written around one material class can still yield a lawful test, provided the interpretation stays close to what the sample represents. The number describes the homogenized submitted sample, and the batch record describes how separate units were produced.

Sample increments make the distinction concrete. A sampler can select material from multiple locations or units according to the applicable plan, creating a laboratory sample with broader batch coverage. The laboratory then reduces and homogenizes that submitted material before analysis. Each stage increases representativeness for a batch-level result, while each stage also combines individual differences. The process is appropriate for compliance and should be described accurately: it estimates the lot through selected material; it does not certify the exact construction of every assembled piece.

The production team can design its own checks around the moments laboratory sampling cannot see. A unit can be weighed before coating, after concentrate pickup and after the outer layer, producing a layer-by-layer record for selected pieces. Recording those observations at the beginning, middle and end of the run reveals drift tied to time. The checks remain separate from official sampling and protect a different decision: whether the assembly process stayed repeatable enough to release the run for regulatory sampling and final packaging.

Homogeneity remains a useful manufacturing goal when it is translated into measurable tolerances. The producer can target repeatable ratios and coverage across individually built units. Regulation can define sampling and release around the batch. Precision comes from keeping those jobs distinct: mixing creates physical uniformity, sampling estimates a population, and process records document assembly. A hand-built composite needs all three concepts, even though only one appears on the certificate.

The Presidential Arizona moon rock collection provides a live product example of the assembled format that these sampling rules must represent.