Diagnose a disturbed isotope clock

A precise calculation can answer the wrong question

At two half-lives, this model has 25 parent units and has produced 75 daughter units. Full retention gives an inferred age of 200 Ma. Now hold elapsed time fixed and reduce daughter retention to 50%. Only 37.5 daughter units remain in the analyzed sample; 37.5 have been exported. Parent stays at 25.

The closed-system calculation sees only the retained sample. It infers 100 × log₂(1 + 37.5/25) ≈ 132.19 Ma. That arithmetic is correct for its inputs, but the full-closure premise is false. The known elapsed age remains 200 Ma. The inference is younger by about 67.81 million years.

Before changing the control, predict how each bar will respond. Then capture 100%, 50% and 0% retention at two half-lives. Identify which quantity stays fixed, which is redistributed and which calculated conclusion changes.

Account for the boundary of the sample

At 50% retention, parent + retained daughter = 25 + 37.5 = 62.5 units. Adding exported daughter restores the original normalized inventory of 100. Missing material is outside the sampled system, not destroyed by the calculation.

The interface knows the initial inventory and exported amount because this is a constructed experiment. A real analyst measuring a mineral does not automatically know its original inventory or all material that left it. Do not treat the export bar as an independently measured field observation.

With 0% retention, the sample contains parent but no radiogenic daughter. The closed-system formula reports zero age at every allowed elapsed time. That does not mean time ran backward, decay stopped or the mineral just formed. It means a ratio that can also occur in a new closed sample was produced by the specified loss mechanism.

Be precise about the disturbance model

The control removes a chosen fraction of accumulated daughter at the selected sampling instant. Parent is never removed. Each point on the age-comparison curve is a separate hypothetical endpoint-loss experiment. The dashed curve is not the time history of a mineral leaking daughter continuously.

A continuous leak, a heating event partway through the history, partial resetting or parent addition would require another model. In particular, removing all daughter long before sampling would allow new daughter to accumulate afterward. The zero-age endpoint here does not describe every mineral that lost daughter in its past.

For a positive elapsed time, reducing retained daughter while holding parent fixed lowers D/P and therefore lowers the inferred age. This is a conclusion about this specified mechanism. It is not a rule that every disturbed radiometric age is too young. Unaccounted initial daughter, added daughter or parent loss can change the inference in different ways.

Compare explanations with additional evidence

Suppose a measured sample returns about 132 Ma. At least two explanations fit the simplified ratio: a closed sample that formed about 132 Ma ago, or an older sample with daughter loss. A single ratio cannot uniquely identify which history occurred.

Create a two-column evidence plan. In one column list observations that could support a coherent closure history, such as agreement among independently selected suitable minerals and consistency with mapped relationships. In the other list observations that would motivate testing disturbance, such as alteration textures, spatial variation associated with altered zones or disagreement between suitable isotope systems. Specify how each observation could challenge your preferred explanation.

Agreement is supporting evidence, not a guarantee that all samples avoided a shared disturbance. Disagreement is a reason to investigate analytical problems, mineral histories and assumptions, not a license to discard the inconvenient date. Preserve individual analyses and sampling context. The browser does not provide a real dataset with which to choose between these explanations.

Connect uncertainty to the inference

The interface gives exact synthetic inventories. Real analyses have uncertainty in isotope ratios, corrections and calibration. An uncertainty interval conditional on full closure would not automatically include systematic error caused by violating closure. Repeating a measurement can reduce some random uncertainty while leaving the same geological-model error intact.

As a numerical sensitivity exercise, suppose P is known as 25 units and retained daughter is bounded between 36 and 39 units. Under the same closed-system formula, the inferred age ranges from 100 log₂(2.44) to 100 log₂(2.56), about 128.69–135.61 Ma. These are input bounds, not a statistical confidence interval. Even this entire interval is below the constructed true age of 200 Ma. A narrow result need not be an accurate reconstruction of history.

Transfer without the sliders

An ideal sample starts with 100 parent units and no daughter. After three half-lives, it loses one third of its accumulated daughter at sampling while retaining all parent. Determine the inventories and predict whether the incorrectly inferred age is above or below three half-lives.

Check the disturbed-clock transfer

Parent is 12.5 units and daughter produced is 87.5. Two thirds of the daughter remain: 58.3333 units; 29.1667 units are exported. The inferred number of half-lives is log₂(1 + 58.3333/12.5) ≈ 2.5025, below the true three. With a 100 Ma half-life this is about 250.25 Ma. The three inventories still sum to 100. The browser's 25-percentage-point control steps do not select this exact retention, so calculate it independently.

Conclude with a conditional claim: state the loss mechanism, show the bias and identify what evidence would be needed to apply that explanation to a real sample.

Reference for closure and geological interpretation: Physical Geology: isotopic dating methods. The endpoint-loss experiment, evidence prompts, sensitivity bounds and transfer case are original teaching constructions.