Mangrove tidal forests have been referred to as the “coastal kidneys” of tropical and sub-tropical estuaries, because of their capacity to filter nutrients, baffle sediments, and provide water quality benefits (Duke, 2011). Worldwide and over the past century, mangrove systems have been under threat from land clearing and/or filling for agriculture, aquaculture, and urbanization purposes (Valiela et al., 2001; Duke et al., 2007; Duke, 2011). Peri-urban mangroves (i.e., adjacent to urban areas) are threatened by elevated contamination from sewage and domestic wastewater inputs (i.e., household water from laundry, dishwashing, and bathing), industrial wastes, storm overflow, and sediment runoff (Bosire et al., 2014; Lugo et al., 2014; Branoff, 2017, 2019; Pérez et al., 2018).

Human activities have greatly increased the flow of reactive nitrogen (N) from land to coastal marine ecosystems, often resulting in coastal eutrophication and causing adverse effects to the structure and function of wetland ecosystems (e.g., Howarth et al., 1996; Deegan et al., 2012; Reis et al., 2017). In a study comparing Puerto Rico rivers draining watersheds ranging from urbanized to more pristine and forested, McDowell et al. (2019) reported that urbanization had caused large increases in the concentration and flux of dissolved N and phosphorus (P) (2 - to 50-fold) in the urban river. In peri-urban mangrove systems, soil fertility can be elevated due to human activities and their associated waste products (e.g., sewage and industrial) (Pérez et al., 2018; Perez et al., 2020). Elevated rates of N and P accumulation were reported in peri-urban mangroves in southeastern Brazil (Sanders et al., 2014; Perez et al., 2020). It is not well understood how N is stored and cycled in peri-urban mangrove systems, although a recent meta-analysis suggests that leaf N content depends on surrounding land cover, with N enrichment in urban systems (Branoff, 2017). Sediment concentrations of heavy metals were also positively correlated with urban development, possibly having some effect on mangrove functions, including N cycling (Branoff, 2017).

The urbanized San Juan Bay Estuary (SJBE) has mangrove systems with varying hydrological alterations (e.g., dredged vs. clogged), levels of urbanization, sewage inputs, and landscape setting (Branoff, 2019; Oczkowski et al., 2020a). We previously found unexpected differences in N storage and cycling in the estuarine subtidal sediments associated with systems of varying degrees of urbanization in the SJBE (Oczkowski et al., 2020b). While subtidal sediments in the most urbanized area (the Caño Martín Peña) contained the most N, they also had the lowest N stable isotope values (Oczkowski et al., 2020a, b). This was surprising because urban runoff is often associated with elevated N isotope values. While the exact mechanisms are still being identified, we hypothesize that N fixation, which is an important source of N to mangroves, may be enhanced in these sediments by the ample supply of carbon (C) inputs from sewage and other urban sources (Oczkowski et al., 2020a, b). These studies also suggested that the San José Lagoon located in the innermost stretch of the SJBE may function as a settling basin and accumulate allochthonous and autochthonous inputs.

San Juan Bay Estuary mangrove forests had soil C burial rates ranging from 88 g m^–2y^–1 at the San José lagoon to 469 g m^–2 y^–1 at the Martin Peña East in recent decades (Wigand et al., 2021). Recent C burial rates were ∼2.5 – 3 times greater than historic decades in mangrove forests at Piñones and Caño Martin Peña (Wigand et al., 2021). Differences in C sequestration rates among the SJBE mangrove systems were attributed to multiple factors: landscape setting, hydrology, vegetation, flooding regime, and anthropogenic inputs (e.g., runoff and sewage). Some changes in mangrove geomorphology and hydrology were associated with rapid urbanization in the SJBE watershed, including canalization and dredging among other disturbances (Lugo et al., 2014; Branoff, 2020a). In the present study, we examine patterns of N storage and accumulation in these same mangrove systems of the urbanized SJBE.

We tested for differences in N accumulation rates among SJBE mangrove sites of varying geomorphology, hydrology, nutrient fertility, and land development. We hypothesized that the mangrove stands in the most developed areas will have the highest sediment %N, N storage, and accumulation rates. However, the drivers of the N cycling and storage are complex and may be modulated by landscape setting, vegetation cover, varying soil characteristics (e.g., salinity, pH, and nutrient content), specific form of N inputs, and hydrology at the sites (Twilley and Day, 1999; Bowen et al., 2020; Perez et al., 2020). Therefore, to assess N dynamics we also reported the dissolved nutrients, salinity, and pH in the porewater, and the %N, molar C/N ratios, and N and sulfur (S) stable isotope ratios in the sediments at each site. We used published sediment accretion rates and bulk densities from the same mangrove sites to report on N storage and accumulation rates in recent (1970 – 2016) and historic (1930 – 1970) decades and estimated watershed-wide N storage and accumulation rates in the SJBE.

The magnitude of N accumulation at both the mangrove site and SJBE watershed-wide scale were compared with population density, food inputs, and associated human wastes to assess the potential socio-ecological links between mangrove N accumulation rates and human-sourced N. The SJBE watershed (area: 41,572 ha) currently has the densest human population on the island, ∼1,850 people km^–2, primarily located in the urban municipalities (i.e., San Juan, Carolina, and Bayamón) associated with the city of San Juan, Puerto Rico (U.S. Census Bureau, 2017). In addition to degree of urbanization, landscape setting, and flushing, we discuss different biogeochemical factors in the mangrove soil and porewater that may influence N storage and accumulation rates. An understanding of N-cycling in peri-urban mangrove systems, often subject to anthropogenic stressors, will support effective land management efforts to sustain mangrove systems and the services they provide.

Human wastewater and other nutrient inputs can enhance the soil fertility and ultimately affect N storage in peri-urban mangroves. In addition, hydrological alterations and land development as well as natural drivers such as varying soil characteristics, salinity, and tidal regime among mangrove systems in the SJBE can affect N storage and accumulation rates. In our study, we measured some of the highest N storage (11.1 Mg ha^–1) and accumulation rates (27.9 g m^–2 y^–1) in the highly urbanized and altered Martin Peña East in recent decades. Mangrove forests at Martin Peña East fringed a poorly drained canal and often received raw sewage, domestic wastewater, and stormwater overflow from the surrounding neighborhoods (Bunch et al., 2000; Oczkowski et al., 2020a). High N storage and accumulation rates were also measured in the least developed Piñones Forest (8.6 Mg ha^–1, 21.2 g m^–2 y^–1, respectively) and the La Torrecilla Lagoon (5.2 Mg ha^–1, 12.0 g m^–2 y^–1, respectively) with a dredged inlet and modified shoreline. Within site variability at La Torrecilla Lagoon was high with one core having significantly lower soil bulk density, sediment accretion rates, and N accumulation rates, suggesting high spatial variability associated with the dredging and shoreline modifications at that site.

While variable across the estuary, N storage in the SJBE falls within the range observed in other mangrove forests. The few published studies that provide direct measurement of N storage in mangrove soils suggest a range of 2.7–56 Mg ha^–1, although the depths (i.e., 20–356 cm) and mangrove types corresponding to these N stocks vary widely (Fujimoto et al., 1999; Alongi et al., 2003; Khan et al., 2007; Bulmer et al., 2016). In the SJBE mangrove soils, N storage in recent decades ranged from 2.1 to 11.1 Mg ha^–1 among sites, similar in magnitude to mangroves: in Rookery Bay (FL) ranging from 2.0 to 4.8 Mg ha^–1, in southern Thailand ranging from 6.1 to 12.9 Mg ha^–1, and in New Zealand averaging 14.6 ± 1.7 Mg ha^–1 (Steyer, 1988; Alongi et al., 2004; Bulmer et al., 2016). The SJBE watershed-wide estimate of 9.4 Mg ha^–1 fell at the lower end of the published range, but just represented 28 cm depth. By extending the trend in N storage to 50 and 100 cm depths for comparison, our SJBE estuary-wide estimates were 13.9 and 20.9 Mg ha^–1, respectively (Figure 6). The estimate extrapolated to 100 cm depth aligns closely with another indirect estimate for global mangrove N storage of 20 Mg ha^–1, although it is not clear what depth range this represents and the authors state this was an unmeasured estimate based on C:N ratios (Reis et al., 2017). Still, the 100 cm extrapolation values in our study ranged from 8.1 Mg ha^–1 at San Jose to 28.6 Mg ha^–1 at Martin Peña East, exemplifying the high spatial variability in N storage of mangrove soils within the same estuarine system.

When considering storage over the past 80 years, the rate of N accumulation in the SJBE was exceptionally high at the peri-urban Martin Peña East site and the most isolated and least urbanized Piñones Forest site. Mean N accumulation rates in the SJBE ranged from 3.0 to 8.6 g m^–2 y^–1 in historic decades and 3.7–27.9 g m^–2 y^–1 in recent decades. The N accumulation rates in recent decades at Piñones (21.3 g m^–2 y^–1) and Martin Peña East (27.9 g m^–2 y^–1) were elevated and ∼2.7–3.7 times greater than historic decades. Nitrogen accumulation rates among riverine and basin forests in the Gulf of Mexico averaged 5.5 g m^–2 y^–1 (Lynch, 1989; Twilley and Day, 1999). The N accumulation rates in recent decades at Martin Peña East and Piñones were ∼4 – 5 times greater than the average N accumulation rates among riverine and basin mangroves in the Gulf of Mexico. However, N accumulation rates in the highly impacted and eutrophic Cubatao Forest and Septiba Bay mangrove systems in southeastern Brazil were as high as 31 and 90 g m^–2 y^–1, respectively; likely because of high urban sewage discharge since the 1900s (Sanders et al., 2014; Perez et al., 2020). The Martin Peña East accumulation rates were similar in magnitude to the Cubatao Forest (Sanders et al., 2014). However, the Septiba Bay N accumulation rates were ∼3 times greater than the Martin Peña East rates and ∼5 times greater than the SJBE watershed-wide N accumulation rate (18 g m^–2 y^–1) in recent decades, suggesting that on the scale of the entire estuary, the SJBE may not be as impacted as other human-modified mangrove systems. When extrapolated down to 50 and 100 cm, SJBE watershed-wide belowground N accumulation rates averaged even lower to 10.9 and 4.8 g m^–2 y^–1, respectively; demonstrating the importance of recent N dynamics, likely anthropogenically driven, to the soil N pool (Figure 6).

We observed different N storage and accumulation rates at the western and eastern ends of the highly developed Caño Martin Peña. All SJBE mangrove sites exist against a backdrop of rising sea levels, with estimated rates for Puerto Rico of ∼2 mm/year (NOAA station #9755371). Biogeomorphic feedbacks between rising water levels and mangrove organic matter production and decomposition results in an accommodation space within which organic matter preservation is enhanced (Gonneea et al., 2019; Rogers et al., 2019). As a result, N accumulation is expected to increase in recent deposits. Anthropogenic nutrient loading and other wetland alterations will further impact the magnitude of N cycling and storage. In the dredged western portion of the canal, the N storage was 67% lower and the N accumulation rates 64% lower than the clogged eastern portion in recent decades. The site differences in N storage and accumulation rates between the western and eastern ends of the canal likely reflect the increased tidal flushing in the Martin Peña West in recent times, which may have lowered particulate deposition and N storage. Generally, mangrove systems that are more flushed conserve less N (Twilley and Day, 1999). In the eastern clogged end of the Caño Martin Peña, the build-up of mangrove litterfall and allochthonous organic inputs (e.g., raw sewage and domestic wastewater) may directly or indirectly contribute to high N storage and accretion rates. Nutrient over-enrichment may promote cyanobacteria and algal build-up in the Martin Peña East, which subsequently would contribute to elevated accretion and N accumulation rates in recent decades. The elevated N accumulation rates in Cubatao Forest and Sepetiba Bay (Brazil) were similarly attributed to nutrient inputs and human impacts associated with rapid urbanization (Sanders et al., 2014; Perez et al., 2020).

It may be argued that in mature, isolated mangrove systems, nutrient recycling may be of greater significance than allochthonous inputs (Twilley and Day, 1999; Alongi et al., 2004; Marchand et al., 2006). Some of the highest measures of N storage and accumulation (this study) as well as C storage were observed in the poorly flushed, least urbanized, Piñones Forest (Wigand et al., 2021). Elevated porewater NH₄⁺ concentrations (mean: 1,057 ± 209 μM in the wet season; 451 ± 76 μM in the dry season), low pH (4.7; Martin et al., 2020), and low denitrification enzyme activity (Oczkowski et al., 2020b) were observed in mangrove soils at Piñones, likely due to high decomposition rates and recycling of carbon and N in the mature system (Marchand et al., 2006).

Historic hydrological alterations contributed to the isolation of the Piñones lagoon and thus its high rates of N accumulation and storage. Prior to the 1950’s the Piñones lagoon likely received water (including agricultural runoff) through a series of drainage and navigation canals from the Río Grande de Loíza and experienced periodic flushing during storms (Webb and Gómez-Gómez, 1998; Cordero, 2015). Following the damming of the Río Grande de Loíza in 1953 (Quiñones et al., 1989), flushing of the Piñones lagoon was greatly diminished, promoting a more isolated mangrove system relying primarily on nutrient recycling for growth and development. Recent N storage and accumulation rates in the Piñones Forest were 242 and 170% greater, respectively, than historical values, which may have been driven by the conservation of N in the peat in recent decades. In addition, cores from Piñones had elevated sediment accretion rates in recent decades (site mean: 5.5 mm y^–1) compared to historic decades (site mean: 3.8 mm y^–1) and the global median rate (2.8 mm y^–1; 95% confidence interval 1.9–3.9) (Wigand et al., 2021). The effect of reduced flushing on N accumulation may have been augmented by expansive and dense plant matter in the mature Piñones Forest that could efficiently trap sediment particles and plant materials.

Although most of the SJBE mangrove soils were reported to contain mangrove root and leaf sources of organic matter, the Piñones mangrove soils may have been supplemented by particulate organic matter from C-4 plants (Wigand et al., 2021). The Piñones mangrove soils were characterized by atypically low C:N ratios (Supplementary Table 2) and an enriched δ¹³C (Wigand et al., 2021). Overland flow of C-4 plant particulates derived from surrounding agricultural lands (e.g., sugarcane), carbon recycling, and elevated porewater salinities may have contributed to the shift to a more enriched δ¹³C at Piñones (Wigand et al., 2021).

The δ¹⁵N values in the Martin Peña East peat were unexpectedly low, generally lower than the δ¹⁵N in the lagoon mangrove peat. We expected a higher δ¹⁵N associated with Martin Peña East due to the high inputs of sewage and human wastewater. However, as reported for the associated subtidal habitats of Martin Peña East, N fixation, possibly associated with common bluegreen algae blooms observed in Martin Peña East, may have caused a shift in the δ¹⁵N profile (Oczkowski et al., 2020b). Compared with atmospheric deposition and N fixation (δ¹⁵N of+2‰ to +8‰), and commercial, inorganic fertilizers (δ¹⁵N of −3‰ to +3‰), N derived from human wastewater (δ¹⁵N of +10‰ to +22‰) is relatively enriched in ¹⁵N (Kreitler et al., 1978; Gormly and Spalding, 1979; Aravena et al., 1993; McClelland et al., 1997).

Porewater N:P molar ratios at the SJBE mangrove sites were often much less than the typical ∼19:1 ratio reported for porewater N:P ratios in the low intertidal zone of mangrove systems (Alongi, 1996). SJBE mangrove forest porewater N:P ratios generally suggested N limitation, with the notable exception of Piñones Lagoon. In Martin Peña East, PO₄⁺ concentrations were very high (>45 μM) and either double (wet season) or nearly equal to (dry season) DIN concentrations. While porewater DIN was not absent (Table 1), the porewater N:P ratios were close to 0, and far from typical mangrove porewater N:P ratios (Alongi, 1996). Given the ponded nature of this stretch of the SJBE and the presumably large urban inputs from the densely settled, unsewered communities adjacent to the site, the relatively low DIN concentration and low porewater N:P ratios are consistent with our hypothesis of high denitrification and N fixation rates in the Martin Peña East (Lenton and Watson, 2000; Oczkowski et al., 2020a, b). The La Torrecilla Lagoon and San José Lagoon also had low porewater N:P ratios (≤10) indicative of N-limitation. Piñones had the highest measured porewater DIN concentrations (driven by high NH₄⁺) of all the sites, resulting in high N:P ratios. In contrast, in examination of the stoichiometry of dissolved N and P in riverine water draining urbanized and more pristine watersheds in Puerto Rico, McDowell et al. (2019) reported little change in the N:P ratios among rivers, although the urbanized river had significant increases in the magnitude of the N and P fluxes.

Porewater salinity was more than twice that of the other locations and pH was low in mangrove soils at Piñones (Table 1). While water outflow is low in Martin Peña East, the region receives substantial amounts of inflow, via urban runoff, as evidenced by documented problems with flooding (Oczkowski et al., 2020a). In contrast, changes in tidal and riverine exchange due to man-induced alterations of the landscape (e.g., dam construction and channelization) minimizes water flow into the Piñones lagoon, except during severe storms. Hydrological alterations may have accelerated natural processes associated with mangrove development, such as more closed mineral cycles and reliance on litterfall and detritus for nutrient regeneration (Lugo and Snedaker, 1974).

The western part of the Caño Martin Peña was dredged during the 1980s as part of a waterway transportation project, which allows a greater influence of tidal waters from the adjacent San Juan Bay (Branoff, 2020a). Greater exchange with San Juan Bay waters may explain the less enriched δ¹⁵N and δ³⁴S in the mangrove sediments at Martin Peña West compared with the clogged Martin Peña East. The lower δ¹⁵N in the flushed Martin Peña West may be due to diluted contributions of allochthonous Caño Martin Peña wastewater N, high water column N-fixation, and/or contributions of marine-sourced N with San Juan Bay tidal inputs. The subtropical and tropical, oligotrophic North Atlantic waters are characterized by isotopically depleted suspended particles often resulting from N-fixation by planktonic diazotrophs (Montoya et al., 2002).

Within the context of the SJBE, with its history of hydrologic modifications (e.g., canalization and dredging; Figure 1) and high porewater salinities, sediment S isotope values generally reflected a combination of two environmental factors: (1) sulfate reduction which contributes to organic matter degradation under anoxic conditions (Canfield, 2001); and (2) the degree of tidal flushing (Oczkowski et al., 2020b). High kinetic fractionation associated with sulfate reduction results in lighter sulfides and enriched sulfates, because ³²S is more quickly reduced than ³⁴S (Brownlow, 1996). Under well flushed tidal conditions, the SJBE sediment is an “open system,” where sulfate from the bay or ocean is continuously replenished, allowing isotopic fractionation resulting in low δ³⁴S as reduced S. In contrast, under poorly flushed or “closed conditions” and in the case of some SJBE sites (i.e., Martin Peña East and Piñones), there is a reduced influx of marine-derived sulfate. High sulfate reduction rates under these closed conditions may cause sulfate reducers to take up more of the less energetically favorable ³⁴S, causing reduced S species to retain a high δ³⁴S (Oczkowski et al., 2020b). Recent genetic sequencing of lagoonal waters at Piñones revealed high representation of sulfur-oxidizing and sulfur-reducing bacteria, which are likely responsible for maintaining S homeostasis in the mostly closed system and prevent the build-up of toxic S compounds in the mangroves (Aviles and Kyndt, 2021).

In the SJBE mangroves we found that δ³⁴S was depleted at depth at most sites, likely reflecting the reduced conditions found at depth in the sediment profile. In addition to reduced conditions contributing to depleted δ³⁴S at depth, a possible source of enriched δ³⁴S in surface sediments at Martin Peña East and San José Lagoon could be sewage discharging into the systems (Oczkowski et al., 2020a, b). Sewage sludge and discharged particles in temperate areas have δ³⁴S ranging from about 0–6‰ (Tucker et al., 1999). The only site that had lower δ³⁴S near the surface than at depth was Martin Peña West, where dredging activities increased tidal flushing with the San Juan Bay (Branoff, 2020a). Generally, we observed a trend toward lower δ³⁴S in surface sediments that received greater flushing (La Torrecilla Lagoon and Martin Peña West) than more closed mangrove systems in the SJBE, supporting the hypothesis that tidal flushing is a factor that controls S isotope composition in mangrove systems, similar to δ³⁴S profiles observed in the SJBE subtidal habitats (Oczkowski et al., 2020b).

Local human pressures on alteration of hydrology, tidal exchange, and soil fertility appeared to cause elevated N accumulation rates in some SJBE mangrove systems (e.g., Martin Peña East and La Torrecilla Lagoon). Anthropogenic stressors (e.g., dredging, filling, and canalization) interacted with natural drivers (e.g., tidal regime and rainfall) that add complexity to predicting nutrient cycling and storage patterns. We measured high mangrove N storage and accumulation in recent decades in Martin Peña East with the clogged canal, in La Torrecilla Lagoon with the dredged inlet, and in Piñones with the least disturbed and expansive forest. Tidal exchange through the dredged inlet at the La Torrecilla Lagoon likely provided a source of particulates and nutrients, which fueled mangrove productivity, N storage, and N accumulation. However, the dredging activities at Torrecilla Lagoon may have also caused the high spatial variability in N storage and accumulation among cores. Recycling and conservation of nutrients in the mature and hydrologically isolated Piñones Forest provided ample N for high mangrove production and N accumulation rates. The highest N accumulation rates (27.9 g m^–2 y^–1) in the SJBE in recent decades were observed in the clogged and eutrophied Caño Martin Peña. This is similar to elevated N accumulation rates in highly eutrophic mangrove systems in southeastern Brazil, apparently due to sewage and wastewater discharges resulting from urbanization since the 1900s (Sanders et al., 2014; Perez et al., 2020). In contrast, atmospheric N inputs from wet and dry deposition were estimated at 0.39 g m^–2 y^–1 in modern times and 0.15 g m^–2 y^–1 in pre-industrial times for the Caribbean Islands (Howarth et al., 1996). Measured wet N deposition in the Luquillo Experimental Forest (1999–2003) was 0.27 g m^–2 y^–1 (Ortiz-Zayas et al., 2006). These estimates suggest that wet and dry atmospheric deposition were generally not a dominant N source to the measured N accumulation rates in the SJBE mangroves.

N accumulation rates in the SJBE mangrove soils ranged from 3.7 to 27.9 g m^–2 y^–1 in recent decades. Mangrove forests accumulate N from natural and anthropogenic watershed sources, tidal inputs, and N fixation. Local anthropogenic stressors may alter N storage and accumulation rates in peri-urban mangrove systems either directly by increasing N soil fertility (e.g., raw sewage and domestic wastewater inputs) or indirectly by altering geomorphology (dredging, filling, and canalization). Other larger-scale external drivers such as rainfall, wind, temperature, and storms may also affect N accumulation in the estuary. All mangrove sites in the urban SJBE, except Martin Peña West, which was dredged in the 1980s, had greater soil %N in recent decades than in historic ones.

The SJBE watershed-wide N accumulation rates in mangrove soils in recent times (0.18 Mg ha^–1 y^–1) were 2-times higher than the rate of food N exported as wastewater (0.08 Mg ha^–1 y^–1), suggesting the potential for mangroves to provide sequestration of human-derived N and act as a coastal kidney. Conservation and effective management of mangrove forests in the Anthropocene are important to the future sustainability of tropical, coastal communities throughout the world because of the many benefits provided: water quality benefits, flood protection, carbon sequestration, as well as food and fiber for millions of people (Valiela et al., 2001; Duke et al., 2007; Lugo et al., 2014; Lovelock, 2020).

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

CW, RM, and AO conceived the study. RM, AH, CW, EH, and AO conducted the field work. ME carried out the radiometric measures of the mangrove soils. CW, RM, and JL carried out the laboratory analyses. BB modeled the estuary-wide N storage. SB modeled the N inputs to the watershed. KM helped with the study design and carried out most statistical analyses. CW wrote the manuscript with input from AO, BB, ME, EW, SB, and KM. All authors contributed to the article and approved the submitted version.

The views expressed in this article are those of the authors and do not necessarily reflect the views or policies of the United States Environmental Protection Agency (EPA). The EPA does not endorse any commercial products, services, or enterprises. Any use of trade, firm or product names is for descriptive purposes only and does not imply endorsement by the United States Government. This report is ORD Tracking Number ORD-041430 and it has been reviewed technically by the United States EPA’s Office of Research and Development, Center for Environmental Measurement and Modeling, Atlantic Coastal Environmental Sciences Division.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.