The countess had risen from the table so as not to alarm her family. By the time she returned, it was already too late to halt the luncheon: her daughters were beginning to experience the very same distress. At the Château de Rasay, on September 6, 1828, a panada served at ten in the morning had introduced peril into the communal meal. A physician was summoned, the afflicted were attended to, and soon the question ran through the entire household: had the harm originated in the kitchen, in a vessel, or from someone who had approached it?
Madame de Saint-Chamans was staying with Count Louis de Saint-Chamans alongside her daughters, Mesdames de Marray and de Tourette. The latter’s two young sons, Henri and Ernest, were likewise present. The adults presently recovered. Henri received the prescribed remedies; Ernest was asleep when the physician gave orders that he should not be disturbed. His condition worsened during the night, and he expired the following evening. The chronicle notes this divergence in care, while offering insufficient grounds to determine whether other measures might have preserved the child’s life. The family had lost one of their own, and the inquiry was bound to begin in the servants' quarters.
The copper saucepans were inspected. No trace of the suspected verdigris was discovered. This initial examination appeared to dismiss a convenient explanation: that of an ill-tended utensil. Yet the cleanliness of the kitchen had been the subject of repeated remarks by Marie Plessé, lady’s maid to Madame de Marray, known in the household as Julie. The cook, the widow Bordier, held a post that Julie was eager to obtain. Suspicion thus assumed the shape of a rivalry over domestic employment.
Marie was soon to wed Gambier, the count’s valet de chambre. The marriage would compel them to depart, as their employers were unwilling to retain them once united. They intended to make for the North, whence Gambier hailed. Yet Julie wished to remain at Rasay as cook. Her future thus hung upon a decision made by others: to marry meant forfeiting her place, while staying at the château required that an office held by another woman fall vacant. For the prosecution, turning the family against the widow Bordier might have paved the way for such a change. This line of reasoning supplied a motive for a dining catastrophe, but it remained to be shown that a desire for employment had translated into a criminal act.
On the day of the meal, Julie laid the blame for the sickness upon the cook’s untidiness. On the morrow, she spoke quite contrarily to the woman, assuring her that her pan was blameless. It was likewise recalled that she had not taken her repast with the other servants, having come downstairs later and lingered near the hearths. Bordier had stepped out and returned several times. The investigation sought in these brief absences the opportunity for an act that no witness claimed to have seen. The mundane comings and goings of the kitchen became the intervals into which a hidden design was inserted.
Upon her arrest, Marie initially denied everything. She subsequently confessed to having introduced an injurious substance into the preparation, while assigning an entirely different purpose to her deed. She maintained that she had intended merely to indispose Ferrand, the footman who carried the dish to the table, out of spite for his unwelcome amorous advances. She believed, she claimed, that the food was meant for him. This account acknowledged a deliberate act, yet disputed both the intended victims and the severity of the outcome. It therefore ran counter to the professional motive that investigators had initially ascribed to her.
Physicians inspected the object from which, by her account, the substance had been drawn. They concluded that the quantity capable of being scraped from it could never account for the violent symptoms observed or for the death of young Ernest. The cook herself had been taken ill after tasting the broth. This partial confession failed to resolve the origin of the harm: either Marie’s explanation was incomplete, or vital elements remained unaccounted for. Justice was confronted with a woman who admitted an intent to do injury, and with consequences that her narrative could not explain.
Yet Ernest’s post-mortem examination revealed no conclusive signs of poisoning. Nor was it proved that Marie had procured any other noxious substance. There was indeed arsenic kept on the estate for domestic purposes; a portion of it lay in an unfastened bureau. The household staff could easily gain access to it. This availability lent weight to the accusation, but it did not demonstrate that Julie had taken any of it. Between the mere presence of a toxic substance in a house and its administration by a designated individual, there remained a chasm that suspicion alone could not bridge.
The preliminary examination then opened a second dossier within the first. Several persons had suspected Marie of being with child that spring. Towards the end of May, she had been taken ill; traces of blood had been observed in her bedchamber and upon her linen. She had offered explanations that were deemed satisfactory at the time. Following the fatal luncheon, these recollections were scrutinized under a new light. Everything in the servant’s past that had once been dismissed now returned to the magistrate’s desk, entangled with an accusation from which its evidence had nevertheless to be kept distinct.
When searches were commanded, Marie confessed that she had been pregnant by Gambier. She stated that she had delivered a stillborn child on May 27, after injuring herself while carrying a heavy tub. She added that she had kept the body for several days before disposing of it. It was discovered in the privies, in an advanced state of decomposition. The medical findings described a fully-term infant, which contradicted the date of conception Marie had asserted. This discrepancy bore upon the duration of the pregnancy; it was not sufficient, in itself, to transform a concealed delivery into proved murder.
The chronicle incorporates further recollections gathered from former masters: suspected pregnancies, and deaths attended by symptoms that some retrospectively likened to poison. Such conjectures cast an aura of sinister repute about Marie without constituting established felonies. Above all, they demonstrate how a single case may draw to itself occurrences that had long remained entirely distinct. The court would have to rule upon two specific indictments: poisoning and infanticide. That both had been examined together did not mean they demanded the same answer.
The trial opened on December 22, 1828, before the Assizes of Indre-et-Loire and spanned three sittings. Marie, initially observed to be composed, was subsequently seized by recurring fainting fits. The old account dwells upon this change of demeanor, which cannot, however, substitute for a rigorous evaluation of the evidence. The witnesses reiterated the substance of their depositions. On December 24, the jury found her not guilty of infanticide, but guilty of poisoning. She was brought back into the courtroom to hear the verdict. She swooned, and the chronicle records that she appeared not to hear the sentence of death.
The original chapter does not recount what became of this condemnation. It leaves a bereaved family, a domestic servant condemned for one offense and acquitted of the other, and a partial confession whose sufficiency the medical experts had challenged. These distinctions must be maintained to the end. The infant discovered in the privies could not be added to Ernest as a judicially established second victim of Marie Plessé. The panada had gathered the family about a single table; the prosecution had gathered two affairs about a single woman. The verdict, for its part, set them apart.