A cloudy week changed the question
In Paris in 1896, Henri Becquerel was testing an attractive possibility: perhaps substances that glowed after illumination also emitted the penetrating rays recently discovered by Röntgen. He put a uranium salt next to a photographic plate wrapped in black paper. Light could not reach the plate, but a penetrating radiation might. After exposure to sunlight, the developed plate carried an image. At first, that seemed to fit the idea of stored light being released in a new form.
Then the weather interfered. Becquerel left prepared plates and their uranium salts in the dark instead of taking them into the sun. When he developed the plates anyway, the images were strong. The result made the original explanation difficult to keep: the salt did not need fresh illumination to produce its rays. Further experiments varied the conditions. The emission continued without the light, heating or electrical excitation that ordinarily makes a material glow.
Sources: [1]
Air became a measuring instrument
A photographic plate gives a visible result, but darkness alone is an awkward ruler. Exposure time, the covering and the photographic process all affect it. The rays offered another way to measure them. They made surrounding air conduct electricity, so a charged body lost its charge more quickly. Radiation knocked electrons loose from atoms in the gas, leaving charged fragments called ions. An electric field could collect these charges as a tiny current.
Marie Curie used this electrical approach to compare materials systematically. The crucial comparison was between chemical forms of uranium. Changing the compound changed many familiar properties, yet the activity followed the amount of uranium present. Something associated with the element itself was at work. Thorium also showed the effect. Radioactivity, the name she gave this property, began to look like a feature of particular atoms.
To compare two samples fairly, the measuring arrangement also had to stay the same. A stronger current could come from more radioactive material, but it could also come from bringing a source closer or letting less radiation be absorbed on the way. Using controlled geometry made the chemical comparisons meaningful. The point was to find a property that followed the material through its transformations. A numerical signal could then distinguish a genuinely more active fraction from a photograph that simply looked darker.
The ore was too active for its uranium
That way of measuring created a new puzzle. Some uranium minerals were more active than their uranium content could explain. A simple possibility was an unknown substance, present in a small amount but emitting very strongly. To find it, Marie and Pierre Curie combined chemistry with the current measurement. After separating an ore into different portions, they asked which portion carried the excess activity. Each measurement decided where the next chemical separation should begin.
The method resembles following a smell through a complicated mixture, except that the guide was a reproducible physical measurement. One active portion behaved chemically like bismuth; another followed barium. In 1898 the Curies announced polonium and, with Gustave Bémont in the radium work, radium. These were difficult claims because intense activity could come from very little material. Chemical concentration, a characteristic optical spectrum and progressively purer preparations supplied different ways to recognize the new substance.
Atoms had an internal source of change
The discovery unsettled the picture of an atom as an unchanging piece of an element. A sample could emit radiation continuously while its ordinary appearance hardly changed. At the time of the first photographic result, the nucleus had not yet been discovered. The modern explanation came later: some nuclei change spontaneously, and the products carry away energy. A long-lived source can seem steady during an experiment because only a tiny fraction of its nuclei changes in that time.
It is useful to keep the sequence of discoveries in view. Becquerel found the spontaneous radiation in 1896. The Curies used the strength of that radiation to search for new elements in 1898. The 1903 Physics prize recognized these contributions together. Pierre Curie delivered his Nobel lecture in 1905; Marie Curie received the separate Chemistry prize in 1911. The first clue was an unexpectedly dark plate. The lasting method was to turn that clue into a measurement and let the measurement guide the search.