Macroscopically, samples from the study area are homogeneous with no clear distinction in hand specimens. The only major difference is that some are vesicular (DBV5, DBV6, and GBV8) and others are dense (GBC1 and GBC2; Figure 3(a)-(b) ). Under the microscope, these lavas have a porphyritic microlithic texture with olivine ( Figure 3(c) ), plagioclase, and diopside phenocrysts.

Olivine occurs as euhedral to subhedral grains, sometimes unhedral, ranging in size from 0.4 to 0.5 mm. Most crystals show destabilization features. These include cracks, kink bands, inclusions of opaque minerals, and reaction crowns. Cracks are observed on crystals of different sizes. Some very cracked crystals occur as aggregates of small associated olivine crystals and sometimes show kink bands. Euhedral crystals are the least cracked and have straight grain boundaries. Reaction crowns concern only a few crystals with a size between 0.4 and 0.1 mm. These have rims with new olivine and plagioclase developing at the expense of the primary olivine, and this is achieved by leaving the rim towards the core ( Figure 3(d) ). Some crystals have a rim corroded by glass containing opaque minerals. At the contact with the other minerals (diopside and plagioclase), no rim reaction is observed.

Diopside is subhedral, less cracked than olivine, and has a size that varies between 0.2 and 0.3 mm. The crystals are mostly elongated sections, with parallel

cleavages forming different angles with the crack plane ( Figure 3(e) ), yellowish in LPNA, with pleochroism ranging from yellow to greenish yellow. Some crystals are poikilitic, containing inclusions of opaque minerals, plagioclase, and olivine, or are sometimes corroded from the edge to the core by glass containing the same opaques. Twins are sometimes observed.

Plagioclase occurs in the form of grains measuring 0.4 to 0.1 mm. It is euhedral to subhedral. It is easily identifiable by its polysynthetic twins, which give it a striped appearance. They are occasionally in contact with diopside, showing a destabilization feature resulting in a corroded core. A type of secondary plagioclase develops, with secondary olivine replacing certain primary olivine crystals.

The mesostasis represents 45 to 55% of the thin section and is composed of diopside, olivine, plagioclase, spinel, titanomagnetite, apatite, and other opaque minerals ( Figure 3(f) ). Titanomagnetite is unhedral, with 20 to 40 µm in size. It is intercalated between the other mineral phases. Some grains with a size not exceeding 100 µm occur as polycrystalline aggregates at the contact of olivine and diopside. The apatite grain is sub-automorphic. It is approximately 30 to 70 µm in size. It is found either as an inclusion in the diopside or intercalated between diopside and olivine.

The major and trace element concentrations of Dibi lavas are presented in Table 2 and Table 3 . The whole-rock SiO₂ content of these lavas ranges between 38.9 and 43.6 wt.%, characteristic of ultrabasic lavas. MgO and Mg-numbers (Mg#) vary from 12.65 to 15.9 wt.% and 67.3 to 71.2, respectively. This Mg# is higher than those described at Wakwa plain (42 - 61 ( Onana François Xavier et al., 2022 )) and Baossi-Warack (40.05 - 64.44 ( Tiabou et al., 2018 )) in the Adamawa Plateau and comparable to the values assumed for primary magmas in some case studies (Mg# > 68; e.g. ( Jung & Masberg, 1998 )). The Al₂O₃ (11.4 - 12.45 wt.%) and CaO (9.57 - 10.5 wt.%) contents are high, with CaO/Al₂O₃ ratios ranging between 0.83 - 0.91. TiO₂ (2.62 - 3.01 wt.%) and P₂O₅ (0.67 - 0.81 wt.%) contents are low. In addition, the studied lavas have relatively medium Na₂O (1.88 - 3.16 wt.%) and low K₂O, varying from 0.85 to 1.83 wt.%, with Na₂O + K₂O contents ranging from 2.73 to 4.99 wt.%. The LOI values are ≤ 1 wt.%, except in one sample with 3.06 wt.% ( Table 2 ).

According to the TAS classification of Le Bas et al. (1986) , the samples are foidites and basanites ( Figure 4(a) ). In the Zr/Ti versus Nb/Y diagram of ( Pearce, 1996 ) ( Figure 4(b) ), the studied samples plot in the alkali basalt field and slightly fall within the transition field between alkali basalt and foidites. The samples plot in the potassic series in the Na₂O versus K₂O diagram after ( Le Maitre, 2002 ) ( Figure 4(c) ), except for one basanite (GBC 2), which falls within the transition field between the potassic and sodic series. Their Na₂O/K₂O ratios vary between 1.73 and 2.21 wt.%. In the major oxides versus SiO₂ diagram ( Figure 5 ), positive trends in Na₂O and K₂O and slight negative trends in MgO, Fe₂O₃, TiO₂, Al₂O_3, and P₂O₅ can be deciphered. These negative trends suggest minor fractionations of olivine, Fe-Ti oxides, plagioclase, and apatite, respectively.

The major element concentrations were also used to calculate the CIPW norm ( Table 2 ). The normative orthoclase has a higher concentration than albite, indicating the potassic nature of these rocks. Olivine (24.52% - 33.52%), nepheline (7.85% - 14.32%), and diopside (17.65% - 26.44%) are present in the Dibi rock samples, along with small amounts of apatite (1.55% - 1.88%), magnetite (2.33% - 2.51%), and ilmenite (4.98% - 5.72%). Based on these CIPW normative calculations and the classification scheme of alkali basalts proposed by ( Chih, 1988 ), the studied lavas are plotted in the alkali picrite basalt field ( Figure 4(d) ), except for the sample GBC 2, which plots in the basanite field.

Primitive mantle normalized ( Sun & McDonough, 1989 ) spider diagram of trace-element abundances for the studied lavas shows that these rocks are enriched in LILEs compared to HFSEs ( Figure 6(b) ). Most of the samples show a slightly positive Nb and Ti anomaly. The plot of REE normalized by chondrite (( Sun & McDonough, 1989 ) Figure 6(a) ) shows an enrichment of LREE over HREEs for the Dibi lavas with La/Yb_CN = 5.15 – 5.88. These lavas do not show any Eu anomaly. The average value of Eu*/Eu is 0.93. All samples show similar trends, suggesting similar REE fractionation degrees and magma source. Their profiles are well correlated with Ocean Island Basalt (OIB) ( Figure 6(a) ) and are clearly distinct from Enriched Mid-Oceanic Ridge Basalt (E-MORB) and Normal Mid-Oceanic Ridge Basalt (N-MORB).

Dibi lavas have high contents of Cr (566 - 796 ppm) and V (227 - 285 ppm).

These Cr and V contents are equal to or greater than the values commonly assumed for primary magmas from peridotite (Cr: 300 - 500 ppm; V: 216 - 295 ppm ( Jung & Masberg, 1998 )). Thorium (6.43 - 7.32 ppm) has a mineral-liquid distribution coefficient of less than 0.01, indicating a strong preference for the melt phase ( Laridhi-Ouazaa , 1989 ). Consequently, thorium becomes highly enriched in the residual melt during magmatic evolution. This element can serve as a differentiation index for variation diagrams ( Figure 7 ). Sc (25.5 - 30.8 ppm), another compatible trace element, has low concentrations ( Table 3 ). The concentrations of incompatible trace elements such as Sr (854 - 949 ppm), Ba (520 - 588 ppm), Zr (301 - 330 ppm), Nb (75.9 - 85.7 ppm), Hf (6.63 - 7.01 ppm), Ta (3.7 - 4.5 ppm), and U (1.65 - 1.88 ppm) are not very variable in these lavas.

Chemical analyses of olivine crystals in the basanites and foidites are presented in Table 4 . They are all chrysolites ( Figure 8(a) ) according to ( Dick, 1989 ); and are magnesian, with a forsterite content varying widely in the foidites (Fo_77.28-87.28) as in the basanites (Fo_78.92-87.04). They are slightly more magnesian than olivine of Ngao Volgar (Fo_78.9 ( Dili-Rake et al., 2022 )) and Baossi Warack (Fo_76-84 ( Tiabou et al., 2018 )) in the Adamawa Plateau, but slightly less magnesian than those of Mount Cameroon (Fo₈₉ ( Wembenyui et al., 2020 )) in the southern part of the CVL. Their Mg# varies from 77.73 to 87.85%. Olivine crystals in Dibi lavas present a wide range of compositions in both the foidites and the basanites. In the foidites, the major element contents of olivine are: MgO (38.75 - 47.20 wt.%), FeO (11.65 - 22.06 wt.%), MnO (0.12 - 0.39 wt.%), NiO (0.13 - 0.38 wt.%), and CaO (0.16 - 0.57 wt.%); in basanites they are: MgO (40.73 - 46.84 wt.%), FeO (12.42 - 19.39 wt.%), MnO (0.16 - 0.39 wt.%), NiO (0.06 - 0.35 wt.%), and CaO (0.17 - 0.42 wt.%). In all these lavas, CaO contents are high (up to 0.57 wt.%). They are different from and higher than those found in the spinel of the mantle xenoliths beneath Adamawa ( Dagwai et al., 2014 ; Nkouandou et al., 2023 ) and some volcanic centers of the Cameroon Volcanic Line ( Tamen et al., 2015 ). Such high CaO contents characterize the olivines of alkaline lavas ( Roeder & Emsly, 1970 ). CaO and MnO contents of olivines are negatively correlated with Mg# ( Figure 8(c)-(d) ). In comparison, NiO contents are positively correlated ( Figure 8(b) ). Similar features were observed in some basalts of the CVL ( Kagou et al., 2001 , Nkouandou et al., 2008 , Fagny et al., 2020 ).

In the conventional Wo-En-Fs diagram ( Morimoto et al., 1988 ), most of the clinopyroxene plots are in the diopside field and a few clinopyroxenes fall into the fassaite field ( Figure 9 ). Clinopyroxene compositions are in the range (Wo₅₁En₃₅Fs₁₄) in foidites and (Wo_50-46.18En_34.8-41.29Fs_10.41-15.92) in basanites. All cpx from Dibi lavas are Ca-rich with Ca/(Ca + Mg + Fe) values ranging from 0.46 to 0.51 (Fe = Mn + Fe²⁺ + Fe³⁺), with an average of 0.49 ( Table 5 ). TiO₂ (0.23 - 4.92 wt.%), Na₂O (0.39 - 0.99 wt.%) and Al₂O₃ (2.10 - 9.43 wt.%) contents are fairly high and varied, indicating the alkali nature of these pyroxenes ( Schweitzer et al., 1979 ). The Fe²⁺/(Fe²⁺ + Fe³⁺) values of clinopyroxenes are relatively high, ranging from 0.45 to 0.80 (with an average value of 0.65), implying that the oxygen fugacity was relatively low ( Wass, 1979 ). The values of the Al^VI/Al^IV ratios are very low in foidites (0.003 - 0.131) and range from (0.035 - 0.863) in basanites. In basanite GBC2, one clinopyroxene crystal shows an Al^VI/Al^IV ratio of 1.067.

Feldspar compositions are illustrated in a ternary plot of the An-Ab-Or ( Figure 10 ). From foidites to basanites, feldspar compositions vary from anorthoclase (solid solution of alkali feldspars) to labradorite (plagioclase). In foidites, the feldspar is typically alkaline (Ab_64.45-65.13An_6.81-9.15Or_26.87-27.08), while in basanites, the alkali type (Ab_60.92-64.11An_8-8.47Or_28.45-32.27) and plagioclase (Ab_37.31-39.67An_57.43-60.70Or_1.99-2.28) are found. Compared to plagioclases ( Table 6 ), anorthoclases show high contents of P₂O₅ (1.24 - 2.12 wt.%) and Cl (0.22 - 0.28 wt.%), while these two elements are almost zero in plagioclases. These alkali feldspars also show higher contents of FeO (6.01 - 7.28 wt.%), MgO (1.05 - 1.26 wt.%), and TiO₂ (1.32 - 2.27 wt.%) than plagioclases: FeO (0.72 - 0.88 wt.%), MgO (0.06 - 0.09 wt.%), and TiO₂ (0.19 - 0.21 wt.%).

Representative chemical analyses and structural formulae of iron titanium oxides are presented in Table 7 . They are present only in basanites. In the sample GBC2, the composition of titanomagnetite is characterized by TiO₂ (20.06 - 23.54 wt.%) contents higher than those of sample DBV6 (TiO₂: 15.17 - 17.06 wt.%). The contents of other oxides vary slightly between these two basanite samples: FeO (62.53 - 64.12 wt.%), MgO (4.41 - 5.61 wt.%), Al₂O₃ (3.88 - 4.58 wt.%) and Cr₂O₃ (0.26 - 1.40 wt.%) in GBC2 and FeO (63.72 - 65.74 wt.%), MgO (4.82 - 5.71 wt.%), Al₂O₃ (5.69 - 7.01 wt.%) and Cr₂O₃ (0.10 - 2.28 wt.%) in DBV6. This titanomagnetite contains up to 0.33 wt.% V₂O₃ and 0.16 wt.% ZnO.

Other mineral phases ( Table 7 ) include spinel (Al-chromite in foidites and Ferrian chromite in basanite, see ( Gawlick et al., 2015 )) and apatite in the basanite GBC2. The Al-chromite oxide contents are as follows: Al₂O₃: 18.74 wt.%, Cr₂O₃: 32.69 wt.%, FeO: 29.07 wt.%, MgO: 12.49 wt.%, TiO₂: 2.79 wt.%, and CaO: 0.04 wt.%. In Ferrian chromite, these contents are Al₂O₃: 9.80 - 14.40 wt.%, Cr₂O₃: 12.78 - 23.37 wt.%, FeO: 39.94 - 53.58 wt.%, MgO: 6.81 - 11.76 wt.%, TiO₂: 5.01 - 9.27 wt.%, and CaO: 0.12 - 0.15 wt.%. Ferrian chromite is therefore more calcic and more titaniferous.

Apatite is characterized by high contents of CaO and P₂O_5, 54.62 and 42.24 wt.%, respectively. Traces of Cl (0.17 wt.%), MgO (0.27 wt.%), and FeO, with content reaching 0.46 wt.%, are present.

In the TiO₂-Na₂O-MnO discrimination diagram of clinopyroxenes ( Figure 11(a) ; ( Nisbet & Pearce, 1977 )), the majority of clinopyroxenes are concentrated in the WPA field, suggesting that the studied lavas are within-plate basalts. This is consistent with the geotectonic scenario of magmatic evolution along the CVL, which implies a pulse of anorogenic alkaline magmatism with an intraplate character favored by major faults that cut the Pan-African crystalline rocks of the Central African Fold Belt ( Njonfang et al., 2013 ).

In the Zr/Y vs. Nb/Y and Nb/Th vs. Zr/Nb ( Figure 11(b)-(c) ) diagrams, the Dibi lavas are compared to those of Wakwa ( Onana François Xavier et al., 2022 ), in the same region (Adamawa), and to those of Bana ( Kuepouo et al., 2006 ), a little further south. The Dibi lavas overlap those studied at Wakwa. They are similar to magmas with an Oceanic Island Basalt (OIB) signature, showing affinity to High-μ (HIMU), Enriched type I (EM1), and recycled components, representing the basalts generated by a mantle plume ( Tsafack et al., 2009 ). All the samples display a uniform pattern on the chondrite-normalized REE ( Figure 6(a) ) resembling that of typical OIB or intraplate alkaline igneous rocks. This indicates that all samples have the same enriched magmatic source.

These findings indicate that the Dibi lavas formed within an intraplate tectonic setting. Keller & Schoene (2018) note that intraplate continental basaltic rocks are predominantly generated through the partial melting of the asthenospheric mantle beneath a substantial continental lithosphere. This interpretation aligns with geophysical research conducted in the Adamawa region ( Poudjom Djomani et al., 1995 ), which also highlights lithospheric thinning and identifies the Moho at approximately twenty kilometers in depth.

A central issue in the study of basaltic magma is determining whether its origin is lithospheric or asthenospheric. Trace element analysis plays a pivotal role in elucidating the source and characteristics of basaltic magmas. For instance, the HFSE/LREE ratio can indicate the parental magma’s source: ratios less than one typically suggest derivation from the lithospheric mantle, while ratios exceeding one are indicative of an asthenospheric mantle origin ( Smith et al., 1999 ; Ali et al., 2013 ). The Dibi lavas exhibit Nb/La (1.30 - 1.51), Zr/La (4.91 - 5.85), and Zr/Ce (2.49 - 2.88) values consistent with an OIB-like asthenospheric mantle source. The elemental ratios (Zr/Nb = 3.76 - 4.02, La/Nb = 0.66 - 0.77, and Ba/Th = 71. 04 - 89.27) further support a HIMU-OIB signature for these lavas ( Weaver, 1991 ). Additionally, the notably low (Rb, U)/Nb ratios (0.22 - 0.55 and 0.019 - 0.023, respectively) align with geochemical characteristics typical of rocks displaying HIMU-OIB signatures ( Kawabata et al., 2011 ).

As magma ascends to the surface, its interaction with surrounding rocks frequently alters its composition. Notably, as illustrated in Figure 2 , in addition to mantle xenoliths, the examined alkaline lavas commonly contain crustal xenoliths. Isotopic analyses typically provide clear evidence of contamination by these crustal materials. However, when such analyses are unavailable, the effects of the traversed rocks can often be inferred from variations in certain trace element concentrations and ratios ( Ali et al., 2013 ; Dili-Rake et al., 2022 ). Accordingly, the discussion of Dibi lava contamination herein is based on its geochemical characteristics.

Referring to studies conducted by Hart et al. (1989) , basaltic magmas affected by crustal contamination exhibit La/Nb ratios > 1.5 and La/Ta > 22. Furthermore, Cai et al. (2010) show that, very often, crustally contaminated mafic rocks show negative Nb and Zr and positive Sr anomalies on the multi-element diagram.

The studied lavas do not show any of these geochemical features ( Figure 6 ), with La/Nb (0.66 - 0.77) and La/Ta (12.82 - 15.29) ratios. In the Rb/Y vs Nb/Y diagram of Cox & Hawkesworth (1985) , Figure 12 , only the basanites GBC2 and GBC1 are positioned within the field of uncontaminated alkali basalts and felsic rocks. The other basanite (DBV6) is not found there (this could be due to metasomatic effects) but is very close to it, as are the foidites, which remain close to the field of uncontaminated alkali basalts and felsic rocks but very far from the Pan-African basement field of Bamenda defined by ( Kamgang et al., 2010 ). All these observations show that the magmas underwent little or no crustal contamination; the presence of mantle xenoliths in many of these lavas indicates that the ascent of the alkaline magma was rapid enough to minimize contamination.

Basalts have long been considered to be the result of the partial melting of peridotite. However, numerous experimental studies have demonstrated that silica-deficient alkali basalts can also originate from the melting of other mantle lithologies such as silica-poor eclogite and garnet pyroxenite, hornblendite, and carbonated peridotite ( Dasgupta et al., 2010 ; Yuan et al., 2020 ).

Recently, some researchers ( Herzberg & Asimow, 2008 ; Yang & Zhou, 2013 ; Yuan et al., 2020 ) have successfully used the MgO, CaO, and Yb contents, Fe/Mn and Dy/Yb ratios, and the FC3MS parameter (FeOt/CaO-3×MgO/SiO₂) to determine the nature of the source lithology of lavas.

To identify the mantle source lithology of Dibi lavas, we used the MgO versus CaO diagram (whole-rock data) from Herzberg & Asimow (2008) ( Figure 13 ). This diagram indicates that the source lithology of Dibi lavas is peridotite.

The relative depletions in HREE and Y presented by the Dibi samples ( Figure 6 ) suggest the presence of garnet in the mantle source. This is because garnet can retain HREE during partial melting ( Goldstein & Francis, 2008 ; Dili-Rake, 2022 ). The studied lavas also present ratios of Y/Yb = 13.54 - 16.01 (> 10) and (Ho/Yb)N = 1.59 - 1.79 (> 1.2), indicative of garnet as a residual phase in the mantle source during the partial melting processes ( Ge et al., 2002 ).

Using the Gd/Yb versus La/Yb diagram from Yokoyama et al. (2007) , the high Gd/Yb ratio indicates more than 8% garnet in the source. Similar garnet content was reported in the source of Baossi-Warack (Adamawa plateau ( Tiabou et al., 2018 )) and the Debunscha maar (south of the CVL at the ocean-continent boundary ( Ngwa et al., 2017 )). Such a large amount of garnet in the source suggests that melting took place within the garnet-stable depth (≥ 80 km ( Niu, 2005 )).

The high Ce_N/Y_N ratios (10.52 - 11.87) suggest a lower degree of partial melting in the generation of Dibi lavas than for Ascension and Bouvet, other OIBs (with Ce_N/Y_N ratios of 4.6 and 3.9, respectively, ( Weaver et al., 1987 )). This partial melting rate is about 3% - 5% ( Figure 14 ).

These degrees of partial melting are higher than the ones inferred at Baossi-Warack (0.5% - 2%) ( Tiabou et al., 2018 ) and Mt. Cameroon (1% - 3%) ( Yokoyama et al., 2007 ) but lower than those of Barombi Koto Volcanic Field (8% - 10%) ( Tamen et al., 2007 ) and Mounts Bangou and Bana (15% - 16%) ( Fosso et al., 2005 ; Kuepouo et al., 2006 ). Dibi’s 3% - 5% partial melting reflects an intermediate mantle process, shallower and more extensive than that of Baossi-Warack and Mt. Cameroon, but not as intense or voluminous as the melting observed at Bangou or Barombi Koto. It suggests a moderate mantle temperature or decompression event, possibly contributing to moderate volcanic activity in terms of volume and explosivity.

Primary magmas in equilibrium with mantle peridotites typically exhibit high Mg numbers (Mg# = 68 - 72) and elevated chromium contents (approximately 1000 ppm) ( Tatsumi et al., 1983 ). The Dibi lavas display Mg# values ranging from 67.3 to 71.2, consistent with derivation from primary mantle-derived melts. However, their Cr concentrations (566 - 796 ppm) are notably lower than expected for primitive melts, suggesting that fractional crystallization has modified the original composition. Geochemical trends observed in major element data support this interpretation. In the Dibi lavas, negative correlations between SiO₂ and MgO, Fe₂O₃-TiO₂, CaO, and P₂O₅ ( Figure 5 ) indicate the fractional removal of olivine, Fe-Ti oxides, clinopyroxene, plagioclase, and apatite. These trends are consistent with mineral fractionation processes typical of alkaline magmatic systems. Additional petrographic and geochemical features reinforce the role of fractional crystallization in the evolution of the Dibi magmas. These include:

Together, these petrographic and geochemical observations indicate that the Dibi alkaline lavas evolved through fractional crystallization of olivine, clinopyroxene, plagioclase, Fe-Ti oxides, and apatite, as is the case for the entire CVL ( Fitton, 1987 ; Yokoyama et al., 2007 ; Kamgang et al., 2013 ; Tiabou et al., 2018 ).

The partitioning of aluminum between the tetrahedral (Al^IV) and octahedral (Al^VI) sites in clinopyroxene is closely linked to the physicochemical conditions under which the mineral crystallized. As originally proposed by Thompson (1947) and Poldervaart & Hess (1951) , aluminum preferentially occupies the tetrahedral site at higher temperatures, while it tends to enter the octahedral site under higher pressure conditions.

In the Al^IV vs. Al^VI diagram ( Figure 15 ), clinopyroxenes from the Dibi alkaline lavas predominantly plot within the igneous field (low pressure) defined by Aoki & Kushiro (1968) and later refined by Aoki & Shiba (1973) . Notably, most clinopyroxene crystals from the basanite sample GBC2 plot in the intermediate field, corresponding to the granulite domain, indicating crystallization conditions distinct from those typical of both eclogitic and primary igneous environments.

These observations suggest that clinopyroxenes from the Dibi lavas crystallized under medium to low-pressure conditions, consistent with rapid magma ascent. The lack of prolonged residence in deep-seated magma chambers allowed crystallization to occur during decompression and cooling at relatively shallow crustal levels, likely during the final stages of magma ascent. This interpretation is further supported by the V_cell vs. V_M1 diagram of Nimis (1999) , which indicates crystallization pressures below 10 kbar ( Figure 16 ). Such pressures are typical of basaltic systems associated with mid-ocean ridges, ocean-island hotspots, or continental

rift settings. Additionally, using the clinopyroxene-liquid geothermobarometer of Neave & Putirka (2017) , equilibrium conditions for the studied clinopyroxenes suggest crystallization temperatures between 981.6˚C and 1051.6˚C, and pressures ranging from 0.9 to 1.6 kbar.

These results collectively support a model of shallow-level crystallization, characteristic of rapidly ascending magmas in intraplate volcanic settings such as the Cameroon Volcanic Line.