Challenges in Identifying Drilled Keshi Pearls
Distinguishing natural pearls from non-bead cultured (NBC) pearls, commonly referred to as “keshi” in the industry, has become increasingly complex. While traditional methods such as radiomicrography were once reliable for the identification of South Sea keshi NBC pearls, recent advances in pearl drilling techniques have rendered these conventional approaches less foolproof. As a result, there is a growing need for advanced data analysis to study the internal structure, chemistry, and other spectral features associated with the mollusk origin to enable reliable separation of natural from NBC pearls.
To further examine challenging structures, GIA’s Mumbai laboratory borrowed pearls represented as natural from a known source. The lot consisted of 40 white to light cream drop and oval pearls (28 undrilled, 7 drilled, and 5 partially drilled). The smallest weighed 0.62 ct and measured 5.17 × 4.16 mm, and the largest weighed 2.95 ct and measured 9.53 × 6.65 mm (see above). The pearls all looked new, with smooth, high-luster surfaces that were free of blemishes. Under 40× magnification, the surfaces showed well-defined overlapping fine aragonite platy structures that accounted for their silky surface sheen.
Energy-dispersive X-ray fluorescence revealed manganese levels ranging from below detection level to 33.80 ppm and higher strontium levels ranging from 1756 ppm to 3364 ppm. Most of the pearls showed an inert reaction under X-ray fluorescence, while some had a weak yellowish green reaction. The latter reaction accounts for the higher manganese levels recorded. The pearls emitted medium to strong yellowish green fluorescence under long-wave ultraviolet light and a similar weaker reaction under short-wave UV. Based on these analyses, a saltwater origin was confirmed for all of the pearls.
Examined with real-time X-ray microradiography (RTX), most of the pearls showed minimal growth arcs, and some revealed faint linear-looking features visible only in certain directions. External and internal features aroused suspicion that these pearls were NBC. Another important observation was related to the size and length of the drill holes. Given the importance of maintaining weight, it is customary to drill drop and oval pearls to approximately one-third of their length. These pearls, however, were drilled to a depth that covered more than two-thirds of their length, while others were fully drilled with oversized drill holes. This indicated the pearls were drilled with the specific intention of concealing internal structures (Summer 2023 Lab Notes, pp. 218–220).
RTX imaging was performed for all the pearls in three directions, with careful alignment along their lengths to detect the presence of any linear feature or its remnants resulting from the drilling process. In most cases, proving a cultured origin only through RTX imaging was challenging. Consequently, the structures were studied using X-ray computed microtomography (μ-CT) analysis, as shown in table 1. In pearls 1 and 2, both undrilled, RTX imaging revealed a few distinct growth arcs, but not much was visible at their centers. Pearl 1 exhibited a small suspicious linear feature along its length, while pearl 2 did not. Interestingly, μ-CT analysis revealed an elongated linear feature running along the length of both pearls, accompanied by a few growth arcs, with the outer nacre lacking many growth features.
Pearls 3 and 4 were partially drilled, and both exhibited drill holes of unusual length and size compared to those observed in natural pearls. RTX imaging showed a few growth arcs toward the outer nacre and minimal growth at the center around the drill hole area. When aligned along their lengths, remnants of a linear feature crossing the drill hole were clearly visible in pearl 3, whereas pearl 4 did not show any evidence of the linear feature. Similarly, μ-CT analysis on both pearls revealed remnants of the linear feature toward the end of their drill holes, indicating an unsuccessful attempt to remove the linear features. Pearl 5, which was fully drilled, had fine growth arcs along its shape in RTX imaging, with a faint shadow feature running parallel to the drill hole. Through μ-CT analysis, a linear feature near the drill hole area was clearly visible.
As observed, structures from the samples closely resembled those observed in South Sea NBC pearls from the Pinctada maxima mollusk (A. Homkrajae et al., “Internal structures of known Pinctada maxima pearls: Cultured pearls from operated marine mollusks,” Fall 2021 G&G, pp. 186–205). In particular, drop and oval pearls from this mollusk should be carefully examined to avoid misidentification as natural due to the absence of structure if only examined under RTX imaging.
Natural drop and oval pearls are highly valued, and those with exceptional appearance and size are not easy to find. In recent decades, some members of the trade have attempted to conceal the linear features visible in South Sea keshi pearls using film or dental X-ray machines and subsequent drilling to remove any evidence. This often results in remnants of the linear features being left behind, as observed in these samples. These remnants can only be identified by carefully aligning the pearls in RTX imaging and through μ-CT analysis. Such subtle features can be easily overlooked during testing, potentially leading to their misinterpretation and misidentification as natural pearls. It is therefore essential for the trade to remain alert and informed about these increasingly frequent practices.
