The road into Kirinyaga was still dark in places from the previous night's rain. Baraka drove himself. He did not want a Crown Ember vehicle making the visit look like a company inspection, and he did not want to arrive like a detective.
Mwangi Njoroge, a local agronomist, met him near an input shop.
“Kifaru Ridge?” Mwangi said. “I know the sign.”
“The sign or the farm?”
Mwangi smiled. “Come and see.”
The place was real: a stone house, a small office structure, a beautiful slope and several rows of healthy coffee trees. It photographed well. But most of the cherries were still green.
“You could selectively pick some red fruit,” Mwangi said. “You cannot get the declared volume from what is here.”
Baraka asked him to use a range, not a dramatic number. They made a rough visible tree count, documented the assumptions and noted that unseen leased land could change the estimate. The report had to be capable of accepting new evidence later.
A neighbor named Kamau came over when he heard them asking about the property.
“Vehicles go in there,” he said. “But I do not see much cherry coming out.”
“What kind of vehicles?” Baraka asked.
“Pickups. Sometimes a small truck.”
“Have you seen coffee bags?”
“Twice. At night. I could not see what was inside.”
Baraka wrote exactly that. He did not promote a neighbor's suspicion into proof.
At a local collection office, Kifaru Ridge appeared on the address list because the company was registered in the area. The seasonal delivery line showed zero kilograms. The clerk warned that private estates could use other mills and sell outside the local route.
“Then zero here does not mean zero production,” Baraka said.
“Correct.”
The clerk identified other mills that could be checked through the buyer's authorization. Baraka added them to the follow-up list.
He also tried the Kifaru office itself. The caretaker said the manager was away and refused access to internal records. Baraka recorded the refusal as a refusal—not as “missing records”—left the authorization and gave the manager a route to respond.
Before leaving, he called Achieng.
“I need to correct another sentence,” he said. “I cannot say the declared harvest was impossible. I can say the visible crop, estimated capacity and local delivery record are inconsistent with the declared volume. If they show leased blocks or outgrower documents, the report changes.”
“That is the report I can defend,” Achieng said.
Mwangi wrote his assumptions beneath the estimate. Baraka photographed crop stages without including private areas of neighboring homes. He did not follow the pickup that entered the Kifaru gate as they left. Kamau's statement was a direction, not a license for private surveillance.
The field visit had not produced a smoking gun. It produced something more useful: a physical source claim that did not fit the land visible under the name, and a list of precise questions that could still be answered.
On the local delivery log, Kifaru Ridge's seasonal record was simple.
Mwangi, the agronomist, made Baraka write the limits of the field estimate into his notes: a roadside visit could miss leased acreage, purchased cherry or outgrower production. Kamau, the neighbor, could describe traffic but not ownership. Baraka accepted each limitation. He wanted a conclusion that could survive somebody proving him partly wrong. The site visit therefore did not say Kifaru produced nothing anywhere; it said the estate story attached to the declared volume was unsupported by the land, crop stage and delivery record they had been shown. The next step had to follow processing, not speculation.
Zero kilograms.