Gem tourmaline is well known from granitic pegmatites in the Nampula and Zambézia provinces of Mozambique (S. Lächelt, Geology and Mineral Resources of Mozambique, Direcção Nacional de Geologia Moçambique, Maputo, 2004, 515 pp.). The Cabo Delgado province in northeastern Mozambique was mentioned by Lächelt (2004) as a pegmatite locality, but to our knowledge no gem tourmaline has been reported from there. However, Farooq Hashmi (Intimate Gems, Jamaica, New York) recently showed us some rough and cut samples of tourmaline that  reportedly came from Cabo Delgado . Mr. Hashmi obtained two parcels of this tourmaline in Dar es Salaam, Tanzania, in mid-2006. One parcel consisted of about 2 kg of broken fragments that ranged from green to orange to brown. Most of the rough was of cabochon quality, with individual pieces averaging 1–2 g. The other parcel contained more-transparent pieces that were waterworn; they averaged 3 g each and ranged from greenish brown to brownish or grayish green. Mr. Hashmi saw several 2–3 kg parcels of the latter material while in Tanzania, and more recently he has seen the rough appear in the New York market. Although he purchased all of these tourmaline samples in Tanzania, he was informed by multiple sources that the material in both parcels came from Cabo Delgado. Mr. Hashmi loaned GIA several fragments and one faceted stone from the first parcel (e.g., figure 1), and one piece of rough and several faceted stones that were cut from the second parcel (e.g., figure 2).

Examination of the cut stone from the first parcel by one of us (EAF)showed the following properties: color—orangy yellow, with very weak colorless to yellow pleochroism; R.I.—no=1.632, ne=1.612; birefringence—0.020; hydrostatic S.G.—3.02; Chelsea filter reaction—none; fluorescence—inert to long-wave UV radiation and weak yellow to short-wave UV; and a weak absorption band at 500 nm visible with the desk-model spectroscope. Microscopic examination revealed numerous negative crystals and lint-like aggregates of colorless inclusions. The several rough pieces from the first parcel (green, greenish yellow, orange, orangy brown, and brown) that were tested briefly showed moderate to strong pleochroism, were inert to long-wave UV radiation, and fluoresced very weak to moderate yellow to short-wave UV. The lighter colored samples had a stronger reaction to short-wave UV. Microscopic examination revealed planes of fluid inclusions and some negative crystals (figure 3), as well as rare colorless low-relief crystals, black disks, and a rounded brassy inclusion (figure 4). A colorless low-relief inclusion in one of the rough pieces was identified as an amphibole by Raman analysis.

The properties reported above are consistent with tourmaline, and the samples were identified as dravite by LAICP-MS chemical analysis performed by Dr. Mike Breeding of the GIA Laboratory in Carlsbad. All of the stones contained minor amounts of Fe and traces of Ti, Ga, and Sr. While the color of some of this Mozambique tourmaline is similar to dravite-uvite from Kenya, the refractive indices reported for the latter are considerably higher (no=1.643– 1.646 and ne=1.622–1.624; see H. A. Hänni et al., “Golden yellow tourmaline of gem quality from Kenya,” Journal of Gemmology, Vol. 27, No. 7, 1981, pp. 437–442). However, the gemological properties of the Mozambique samples are comparable to those documented in samples with a similar hue range from Morogoro, Tanzania (Summer 2006 Gem News International, pp. 182-183). LA-ICP-MS analysis of the cut stones and rough sample from the second parcel (again, see figure 2) showed that they consisted of either dravite or, surprisingly, liddicoatite. The three dravites were darker than most of the six liddicoatites that were tested.

The dravite samples from the second parcel showed the following properties: color—dark brownish green to dark grayish green, with moderate green-to-colorless pleochroism; R.I.—no=1.640, ne=1.620; birefringence—0.020; hydrostatic S.G.—3.04–3.06; Chelsea filter reaction—none; fluorescence—inert to long- and short-wave UV radiation; and an absorption band at 500 nm visible with the desk-model spectroscope. Microscopic examination revealed only a very small “fingerprint” in one sample; no other inclusions or growth features were seen. LAICP- MS analysis showed that along with the elements normally associated with tourmaline, these dravite samples contained traces of Ti, Sr, and V.

The liddicoatite samples had the following properties: color—light brownish green to dark  greenish brown, with weak to moderate green-to-colorless pleochroism; R.I.—no=1.640, ne=1.620; birefringence—0.020; hydrostatic S.G.—3.06; Chelsea filter reaction—none; fluorescence—inert to long- and short-wave UV radiation; and a weak absorption band at 500 nm visible with the desk-model spectroscope. In general, the stones were of high clarity; microscopic examination revealed angular color zoning, small transparent crystals, and growth tubes. Two of the stones contained a distinct dark green triangular color zone. The gemological properties are typical of liddicoatite (see D. M. Dirlam et al., “Liddicoatite tourmaline from Anjanabonoina, Madagascar,” Spring 2002 Gems & Gemology, pp. 28-53), but the homogeneous brownish green to greenish brown coloration is rather unusual for gem-quality liddicoatite (see, e.g., Winter 2002 Gem News International, pp. 356–357). We could find only one literature reference to liddicoatite from Mozambique: a chemical analysis of a sample from Muiane, Zambézia Province (T. G. Sahama et al., “On tourmaline,” Lithos, Vol. 12, pp. 109–114). Gem dravite and liddicoatite are not geologically associated with one another in nature, so it is likely that the samples from that alluvial parcel originally came from different deposits in Mozambique. This is also suggested by the very different trace-element signature for the liddicoatite, as measured by LA-ICP-MS: Ti, Mn, Ga, Sr, Pb, and Bi.

Eric A. Fritz, Brendan M. Laurs, and John I. Koivula
GIA, Carlsbad