A comparison of Herring Gull Larus argentatus and Lesser Black-backed Gull Larus fuscus nest sites: their characteristics and relationships with breeding success

Nest sites used by Herring Gulls and Lesser Black-backed Gulls in a mixed colony were compared. Correlations between features surrounding a nest and breeding success at different stages of the breeding cycle are presented. Lesser Black-backed Gulls nested on more vegetated and flatter areas than did Herring Gulls, even though the latter species had a lower hatching success at less vegetated sites. The difference in the general topography of nest sites between the species suggests that the Lesser Black-backed Gull, through an individual defence strategy, may be better adapted to use sites which are accessible to ground predators than the Herring Gull. The fledging success of Lesser Black-backed Gulls tended to increase with an increased proportion of relatively tall vegetation close to the nest. The Lesser Black-backed Gull may therefore be able to leave its young unattended and rely on their concealment for protection against predators. In contrast, the Herring Gull may rely more on parental vigilance to protect young. More frequent attendance by adult Herring Gulls at the nest site during chick rearing compared with Lesser Black-backed Gulls supports this hypothesis.     

SPECIES ISOLATION BETWEEN THE HERRING GULL LARUS ARGENTATUS AND LESSER BUCK-BACKED GULL L. FUSCUS

There is a large, mixed colony of the two “ring” species, the Herring Gull Larus argentatus and Lesser Black-backed Gull L. fuscus, on Walney Island, northwest Lancashire. These birds are nesting at the very high density of one nest/40 square yards, or more, but although they defend their territories against both species indiscriminately, there is effectively no hybridization. This paper discusses the nature of the species isolation mechanism, and its function. Since the two species can produce fertile hybrids, the mechanism must be of an ecological/behavioural nature, rather than morphological incompatibility. It is shown that there are slight differences in breeding season and habitat, but these do not seem to be great enough to account for the high degree of isolation. It is likely that species isolation depends primarily on the female's choice of a mate. It is suggested that, as specific cues, she uses the differences in call-note tones, and the colour of the back (and perhaps also of the eye-ring), or both. Herring Gulls and Lesser Black-backs are adapted to slightly different niches. The overlap is so great, however, that any hybrid is unlikely to be at a disadvantage; but the overlap can only have arisen very recently, as a result of both species taking advantage of the increased availability of human refuse. It is possible that the isolation mechanism was evolved to cope with earlier conditions, when food was more limited, and the species' niches more sharply defined.     

Timing of primary moult in adult Herring Gulls and Lesser Black-backed Gulls

Summary Timing of primary moult in relation to the breeding cycle is presented for 113 actively moulting adult Herring Gulls and 79 Lesser Black-backed Gulls from Walney Island, England. Moult in both species occurred about when the eggs hatched in mid-May. The entire Herring Gull population began to moult the primaries within a period of 50 days. Lesser Black-backed Gulls started to moult 10 days later than Herring Gulls but many birds were not yet moulting as late as 5. August when I left the study area. It is suggested that the population of Lesser Black-backed Gulls consists of residents and migrants and that the former begin to moult earlier than the latter.

---

Der zeitliche Ablauf der Handschwingenmauser von Silber- und Heringsmöwe

Zusammenfassung Bei 133 ad. Silber- und 79 ad. Heringsmöwen von Walney Island, England, wurde der Beginn der Handschwingenmauser in Beziehung zum Brutzyklus untersucht. Bei beiden Arten setzte die Mauser ungefähr Mitte Mai zum Zeitpunkt des Schlüpfens der Jungen ein. Alte Silbermöwen begannen innerhalb von 50 Tagen mit der Handschwingenmauser. Die Heringsmöwe fing etwa 10 Tage später an, doch hatten manche Individuen noch bis Anfang August nicht mit der Mauser begonnen. Die Heringsmöwen-Population auf Walney besteht vermutlich aus Stand- und Zugvögeln; erstere scheinen früher zu mausern.

     

THE EFFECTS OF DIURNAL AND TIDAL PERIODICITIES IN THE NUMBERS AND ACTIVITIES OF HERRING GULLS LARUS ARGENTATUS IN A COLONY

The average numbers of Herring Gulls Larus argentatus present in a breeding colony on Walney Island, Cumbria, were found to vary with the tidal cycle but to remain effectively constant with time of day through the breeding season. An activity survey, based on 50 Herring Gulls observed at half-hourly intervals during March and April 1973, showed that sleep and rest varied inversely with each other with sleep increasing to 50 per cent at midday. After a peak in the proportion of gulls asleep four hours before low tide, sleeping progressively decreased until low tide; seemingly a result of resident gulls waking and remaining more alert as others left the colony in search of food. Preening was constant throughout the day and tide cycle. Other behaviours (mostly courtship and agonistic behaviour associated with territory defence) increased slightly during low tide and were more common early and late in the day. Night observations of the gulls' activities showed that there was a peak of sleeping between midnight and 02.00 hours. It is suggested that Herring Gulls have a bimodal diel sleep pattern.     

SOCIAL STIMULATION OF REPRODUCTION IN HERRING AND LESSER BLACK-BACKED GULLS

The date of the first egg laid and the position of nests were recorded for all pairs in 10 study areas within the mixed Herring and Lesser Black-backed Gull colony on the southern end of Walney Island, Lancashire. The study areas differed from one another in density of breeding birds and in proportion of each species. These data were collected in order to determine if evidence could be found for synchrony of reproduction by social stimulation (the "Fraser Darling Effect"). The gulls in different study areas showed considerable variability in onset and spread of laying; however, attempts to correlate a number of variables, such as density of breeding birds in each area with median date or spread of laying, average nearest neighbour distance with median date of egg-laying, and median laying date for one species with that of the other species in the same section of the gullery, failed to produce convincing evidence for inter- or intraspeciiic social stimulation of reproduction.     

Reproduction and survival of Glaucous Gulls breeding in an Arctic seabird colony

ABSTRACT.   Despite being widespread and easily observed, little is known about the life history of Glaucous Gulls ( Larus hyperboreus ). From 1984 to 2007, we examined their breeding biology and demography at Coats Island, Nunavut, Canada, where they nest alongside a colony of 30,000 pairs of Thick-billed Murres ( Uria lomvia ). The gulls fed mainly on murre eggs and chicks and by scavenging adult carcasses. The median age at first breeding was 5 yr, and the mean age was 4.8 ± 0.9 yr. Adult survival was estimated as 0.84 ± 0.03 (SE). The mean clutch size was 2.56 eggs and the mean number of young reared per year was 1.6 (range = 0.9–2.2). Birds reared at the colony provided 40% of recruits. Assuming that survival of locally reared chicks that emigrated was similar to that of chicks that returned to the colony, about 22% of the young gulls survived to breeding age. The timing of breeding by Glaucous Gulls appeared related to the timing of laying by murres. Although the demographic characteristics of Glaucous Gulls in our study were similar to those of populations of other large gulls, adult survival was at the lower end of the range for populations of large Larus gulls. There is some evidence that Glaucous Gulls exhibit lower survival than large gulls breeding in temperate areas, possibly because of contaminant burdens. In general, however, the demographic characteristics of large gulls show little variation and are probably a product of their common phylogeny.     

Plenty of rooftops with few neighbours occupied by young breeding Yellow-legged Gulls (Larus michahellis): Does this occur at the expense of their health condition?

Gulls, as largely flexible opportunistic individuals, have been increasingly breeding in many cities around the world, but it is still unclear whether urban habitats are of equal or higher quality than traditional natural habitats or represent an ecological trap with immediate reproductive benefits but longer-term detrimental consequences to health. Here we present a study of breeding parameters (nest density, egg dimensions, clutch size, hatching success and adult body condition) and physiological parameters (erythrocyte sedimentation rate, heterophil/lymphocyte ratio, haemoglobin concentration and measurements of oxidative stress) as indicators of the general health condition of Yellow-legged Gull Larus michahellis adults and chicks from natural and urban colonies. Yellow-legged Gulls in the largest urban area (Porto) laid smaller eggs and clutches, showed a significantly lower occurrence of inflammatory processes in chicks, and showed a slower early chick growth than in the natural colony of Deserta. This suggests that urban gulls might be facing important trade-offs between the advantages of breeding in lower density urban colonies, with fewer intraspecific interactions and a lower disease transmission probability, and the disadvantages of having an anthropogenic diet usually lower in nutritional value.     

Wie richtet eine Flußseeschwalbenkolonie (Sterna hirundo) ihr Abwehrverhalten auf den Feinddruck durch Silbermöwen (Larus argentatus) ein?

How a Common Tern (Sterna hirundo) Colony Defends itself against Herring Gulls (Larus argentatus) The subject of this study is the anti-predator behaviour of a small common tern colony near a large herring gull colony on the island of Mellum, West Germany (Fig. 1). In 1980 the number of gulls crossing this tern colony increased during the terns' chick-stage (Fig. 4). Observed predation of tern chicks was independent of tide and time of day (3., Fig. 5). The frequency of tern reactions corresponded to the number of herring gull crossings (Fig. 5, Table 1). The terns' responses increased between morning and evening (Fig. 8). Tern up-flights and attacks increased absolutely and as a percentage, with the advance of the breeding season (Fig. 3, 4). They were positively correlated with the observed chick predation and the number of pairs with chicks, most markedly with chicks older than 5 days (Figs. 3, 4; Table 1). This increased defence was maintained by fewer pairs as, by then, many had lost their own broods (Fig. 4). As the breeding season progressed, herring gulls increasingly became the main cause of tern up-flights and the object of the attacks (Figs. 9–11). The up-flights of the whole colony, which occurred frequently and spontaneously during incubation, were observed only rarely after hatching and were almost exclusively a response to herring gulls (Figs. 10, 12). The lower herring gulls flew over the colony, the more frequently common terns flew up or attacked and the more individuals were involved in these responses (Figs. 6, 13, 14). During the breeding period, communal up-flights and attacks by terns increased as a percentage (Figs. 12, 13, 15–17). Group-attacks effected changes in the gulls' flying-routes more often than did individual attacks (Fig. 18). Despite the defence behaviour and its adaptation to the predation pressure, herring gulls often succeeded in robbing chicks. This is why the breeding success of the common tern was poor (< 0.4 chicks/nest). Possible reasons for this are discussed.     

Nest predation and nest site selection among Eiders Somateria mollissima: the influence of gulls

Hooded Crows Corvus cornix, Great Black-backed Gulls Larus marinus and Herring Gulls L. argentatus were the main nest predators in an Eider population in southwest Sweden. The clutch sizes of Eider nests within gull colonies did not differ from those outside gull colonies. The proportion of Eider nests destroyed by predators was significantly lower within than outside gull colonies, especially on islands with Lesser Black-backed Gulls L. fuscus. Although the difference was not significant, the survival time of simulated Eider nests was higher within than outside gull colonies. On Eider islands with gull colonies, foraying crows spent more time within the colony area than expected by chance. However, crows apparently avoided an area around each gull nest, and we suggest that the colonies, to some extent, protected Eider nests against predation. The density of Eider nests was higher on gull islands than on gull- free islands, and higher within than outside the gull colonies. However, the association with gulls was weak compared to that displayed by some other waterfowl.     

Effects of cleaning a tidal river of sewage on gull numbers: a before-and-after study of the River Tyne,northeast England

For most of the 20th century, the tidal reaches of the River Tyne in northeast England were used as an open sewer for organic wastes from almost a million people. This produced anaerobic conditions in the river, but supplied an appreciable source of food for five species of gulls. Observations on the number of gulls using the river were made in 1969/70 to establish the situation before a sewage treatment scheme was implemented. By 1994, the volume of untreated sewage discharged into the river had been reduced by 97% and by 86% in the study area, and a second survey was carried out to evaluate the effect of the clean-up on gull numbers. Overall, gull numbers showed a significant decrease of 37% between the two study periods. The decrease in winter was 38% and was also significant, while the decline in summer was 33%, but not significant. Considering individual species, numbers of Common and Great Black-backed Gulls declined by 93% and 91% respectively. In both these cases this is presumed to be due to the decline in available food. Black-headed Gull numbers did not change significantly and it appears they have been able to find alternative feeding sites such as mud flats exposed at low tide and the outfall where treated effluent is discharged. Numbers of Lesser Black-backed Gulls and Kittiwakes rose dramatically as a result of increases in breeding numbers in the area. Neither species feeds along the river. Numbers of Herring Gulls did not change significantly — it is likely that the effect of the decline in food availability was masked by the increase in breeding numbers in towns along the river banks. The different response of each of the species of gulls indicates the differences in their ecology and breeding status in the area.     

MIGRATION AND DISPERSAL OF AUDOUIN'S GULL LARUS AUDOUINII FROM THE EBRO DELTA COLONY

Oro, D. & Martinez, A. 1994. Migration and dispersal of Audouin's Gull Larus audouinii from the Ebro Delta colony. Ostrich 65:225-230.

Migratory paths and dispersal patterns of Audouin's Gulls (Larus audouinii) born in the Ebro Delta have been studied, as well as their degree of philopatry. The gulls always disperse southwards from the colony, and only the juveniles seem to disperse in other directions. Dispersal patterns of 2y, 3y and older gulls are similar, wintering mainly on Iberian Mediterranean and Moroccan Atlantic coasts. Juveniles migrate before the rest of the age classes and they winter south of older gulls, in the Senegambia region. However, conclusions on Audouin's Gulls movements are difficult to draw, owing to the large differences in resighting and recovery efforts. Data presented also suggest that Audouin's Gulls do not appear to breed before the age of four years.     

Habitat selection and breeding success in Yellow-legged Gulls Larus cachinnans

The habitat selection and breeding performance of Yellow-legged Gulls Larus cachinnans were studied in the Medes Islands colony, northeastern Spain, during 1995 and 1996. Of the three main habitats on the islands (shrubs, grass and bare areas), gulls first occupied those with the highest percentage of tall vegetation. Gulls tended to select nest sites with 20–75% cover despite great differences in the cover in the habitats and territories, suggesting that the presence of a suitable nest site may play a major role in the choice of breeding habitat. Nest-site tenacity did not influence the preferences of gulls at any level since the same pattern of choice was observed in an area subjected to annual culls (i.e. where most of the breeding pairs were culled annually and replaced by naive birds). In spite of great differences in the physical characteristics of the habitats, little difference was found in breeding performance of the gulls between habitats. Gulls nesting in the least preferred habitat (i.e. mainly bare) had smaller clutches than those nesting in the other two habitats, possibly as a result of their later seasonal laying. Despite the similar breeding success in different habitats, gulls did not seem to distribute according to the ideal free model reported for Herring Gulls Larus argentatus since the density in the preferred habitat (i.e. shrubs) was never higher than in the other two. We suggest that the habitat selection by Yellow-legged Gulls within the colony could follow an ideal despotic distribution.     

Summer in the city: behaviour of large gulls visiting an urban area during the breeding season

Capsule: Large numbers of Herring Gulls Larus argentatus and Lesser Black-backed Gulls Larus fuscus from a traditional colony in the Netherlands visit an urban area for food in the chick rearing period, causing nuisance while doing so.

Aims: To assess the potential contribution of large gulls breeding in a traditional colony to gull–human conflicts in cities.

Methods: Colour-ringed gulls from a nearby colony were counted in the study area throughout the 2014 breeding season. The average numbers observed daily per species and sex were compared between different breeding phases.

Results: Fluctuations in numbers of both species could be explained by breeding phase. Numbers of females of both species and male Lesser Black-backed Gulls dropped significantly during laying and incubation. Numbers peaked post-hatching, coinciding with increased engagement in nuisance events and more frequent displacements within the study area.

Conclusion: Large gulls from a traditional colony frequently visited the urban study area, especially when food demand for chick provisioning was high, suggesting that city visits are motivated by accessibility of urban food. The proportion of rooftop breeding gulls in this area is low. Therefore, measures to avoid nuisance should focus on reducing food availability rather than controlling gull breeding.     

Indexing winter gull numbers in Great Britain using data from the 1953 to 2004 Winter Gull Roost Surveys

Capsule Winter Gull Roost Survey data spanning 50 years were used to generate population indices.

Aims To evaluate how wintering numbers of five gull species have changed in Great Britain over the last five decades.

Methods Generalized linear models were used to relate gull numbers to habitat, site and year factors, and so derive species‐specific indices for nine regions of Great Britain. Regional models considered data from different timescales depending on coverage.

Results Patterns of change varied by species and region. All species showed increases in number over the period 1953 to 2004. In most regions, Black‐headed Gull Chroicocephalus ridibundus numbers have declined since peaks between 1973 and 1993; Common Gulls Larus canus have also declined recently in some regions. Lesser Black‐backed Gull L. fuscus numbers have increased dramatically since 1953, whereas numbers of Herring Gull L. argentatus showed large declines between 1963 and 1983. Great Black‐backed Gull L. marinus numbers have increased in the west and the Midlands, but recently declined in eastern regions.

Conclusions Numbers of wintering gulls in Great Britain have shown rapid changes over the last five decades, reflecting changes in the sizes of breeding populations. These changes are likely to be associated with changes in human activities and resource availability.     

Genetic Affinities Within the Herring Gull Larus argentatus Assemblage Revealed by AFLP Genotyping

To date, the taxonomic status of circumpolar breeding populations of the Herring Gull Larus argentatus, the Lesser Black-backed Gull Larus fuscus, and the closely related Yellow-legged Gull Larus cachinnans has been based on differences or similarities in phenotype, morphology, and feeding and premating behavior. To shed some new light on the many taxonomic uncertainties surrounding these taxa, we describe the results of a large DNA study based on comparing the distribution of 209 biallelic markers among 109 gulls, representing 11 gull taxa of the Herring Gull assemblage and the Common Gull Larus canus. A detailed phylogenetic analysis failed to show clustering of individuals into groups representing either geographic origin or phenotype. Alternatively, birds were grouped into taxa defined on the basis of phenotype and geographic origin or phenotype alone. Genetic analyses revealed significantly different genetic distances between all pairs of taxa. However, based on these genetic distances, again no consistent phylogenetic tree could be constructed. Analysis of molecular variance indicated that about 77% of the total genetic variability among these gulls could be explained by within-taxon differences. Only 23% of the total genetic variability was due to genetic differences between taxa, irrespective of their species or subspecies status. Although this seems to challenge the current taxonomic treatment of the herring gull assemblage, our results are too premature and too incomplete to recommend a drastic change. Received: 27 January 2000 / Accepted: 18 September 2000     

Expansion of the Mediterranean Gull Larus melanocephalus in Poland

Since 1981, when the first breeding pair of Mediterranean Gull Larus melanocephalus was recorded in Poland, the population of this gull has increased considerably. Its population size was stable until 1997, not exceeding ten pairs annually; thereafter, an increasing number of sightings were made, and during the last 5 years between 26 and 39 breeding pairs have been recorded in Poland. To date, breeding sites have been established in a total of 45 sites (maximum of 19 sites in a given year), with 27% of these (43% of all broods) found on islands located in the middle course of the Vistula River. Breeding sites have also included artificial reservoirs, such as dam reservoirs (20% of sites, 19% of broods), gravel pits (9% of sites, 13% of broods) and fishponds (24% of places, 8% of broods). Mediterranean Gulls were found to nest only within the colonies of other Laridae. Single pairs were recorded at 40% of the breeding sites, whereas a maximum of two to five pairs were recorded at 47% of the other sites. The great majority of breeding attempts were recorded in Black-headed Gull L. ridibundus colonies. In two cases, Mediterranean Gulls bred within mono-specific colonies of Common Gulls L. canus consisting of 60–150 pairs. The biggest concentrations of breeding sites were along the middle course of the Vistula river and in the southern part of Poland.     

Intercolony variations in movement patterns and foraging behaviors among herring gulls (Larus argentatus) breeding in the eastern Wadden Sea

Herring gulls (Larus argentatus) are opportunistic predators that prefer to forage in the intertidal zone, but an increasing degree of terrestrial foraging has recently been observed. We therefore aimed to analyze the factors influencing foraging behavior and diet composition in the German Wadden Sea. Gulls from three breeding colonies on islands at different distances from the mainland were equipped with GPS data loggers during the incubation seasons in 2012–2015. Logger data were analyzed for 37 individuals, including 1,115 foraging trips. Herring gulls breeding on the island furthest from the mainland had shorter trips (mean total distance = 12.3 km; mean maximum distance = 4.2 km) and preferred to feed on the tidal flats close to the colony, mainly feeding on common cockles (Cerastoderma edule) and shore crabs (Carcinus maenas). In contrast, herring gulls breeding close to the mainland carried out trips with a mean total distance of 26.7 km (mean maximum distance = 9.2 km). These gulls fed on the neobiotic razor clams (Ensis leei) in the intertidal zone, and a larger proportion of time was spent in distant terrestrial habitats on the mainland, feeding on earthworms. δ¹³C and δ¹⁵N values were higher at the colony furthest from the mainland and confirmed a geographical gradient in foraging strategy. Analyses of logger data, pellets, and stable isotopes revealed that herring gulls preferred to forage in intertidal habitats close to the breeding colony, but shifted to terrestrial habitats on the mainland as the tide rose and during the daytime. Reduced prey availability in the vicinity of the breeding colony might force herring gulls to switch to feed on razor clams in the intertidal zone or to use distant terrestrial habitats. Herring gulls may thus act as an indicator for the state of the intertidal system close to their breeding colony.     

Experimental evidence for the relationship between food supply, parental effort and chick survival in the Lesser Black-backed Gull Larus fuscus

We studied breeding success, chick growth, parental effort and chick behaviour in two groups of Lesser Black-backed Gulls Larus fuscus whose chicks were provided with additional food until 7 days after hatching or until fledging. These data were compared with those from control pairs which we studied simultaneously to test the hypotheses that food was in short supply during the chick stage at the colony site and that in such circumstances the behaviour of adults and young is mainly responsible for the low success. Pairs whose chicks were fed with additional food until fledging showed a higher fledging success than control pairs (intermediate for pairs of first experimental group). During the first week after hatching, experimental adults of both groups were present together at the territory for longer than control pairs. In adult females of experimental pairs, the length of feeding trips was shorter than in females of control pairs, whilst the rate of chick feeding was more frequent in the experimental broods. After the chicks were 7 days old, differences were significant only for the experimental pairs whose chicks were provided with additional food until fledging. Chicks fed until fledging showed a higher daily mass and wing-length increments and reached a higher fledging mass at an earlier age than both control chicks and chicks which were provided with additional food until day 7. Starvation occurred only in control chicks and in chicks of the first experimental group after we had stopped providing food. When food was in short supply, fledging success of gulls was adversely affected as a result of both starvation (because of the lower feeding rates of chicks) and a higher predation rate (arising from changes in behaviour of both adults and chicks).     

The feeding and breeding ecology of Barn Owls Tyto alba in Peninsular Malaysia

Barn Owls have only recently colonized Peninsular Malaysia, nesting in the roof spaces of houses in oil palm estates and feeding on the rats which inhabit these plantations. Pellet analysis showed that the prey spectrum was confined almost entirely to three species of the genus Rattus which are the major pests of oil palm. There was no annual variation in diet. Breeding showed a broad seasonality but occurred in all months of the year. Mean clutch and brood sizes of 6.6 and 4.6 respectively were recorded, most pairs producing two broods a year although on two occasions three were raised. Overall hatching success was 69.0% with first clutches more successful (79.9%) than second (57.3%). First broods fledged 86.1% and second broods 69.1% of young fledged. Comparison of growth rates of different sized broods suggested that there is a physiological maximum at which all broods proceed irrespective of brood size. The behaviour al changes needed in hunting techniques when colonizing dense plantations rather than the more usual open habitat of Barn Owls is discussed. The breeding strategy seems to be one of producing large clutches and broods, and frequent breeding attempts in a habitat with a high potential carrying capacity.     

Sex-biased mortality in the Lesser Black-backed Gull Larus fuscus during the nestling stage

Theory predicts that skewed progeny sex ratios should be relatively common in vertebrate populations. In most birds this has proved hard to substantiate due to the difficulties associated with identifying the sex of large samples of chicks. This study reports the success of a new molecular DNA technique in determining the sex of 601 newly-hatched Lesser Black-backed Gulls Larus fuscus
There was no evidence of any adaptive sex ratio within broods. Male chicks were found to be disproportionately large and to grow at a faster rate than females. The overall sex ratio changed significantly from 0.484 (male/male + female) at hatching to 0.399 by fledging, probably due to male susceptibility to starvation. Mortality also increased significantly with hatching order, an effect often observed in species like the Lesser Black-backed Gull where hatching is asynchronous. I discuss the possibility that hatching asynchrony may in fact be a strategy employed to prevent excessive skews developing in progeny sex ratio whenever variable differential mortality is likely.
The results appear to vindicate Fisher's (1930) hypothesis which predicts the overproduction of the 'cheaper' sex. However, as the skewed sex ratio may be determined more by unpredictable environmental factors, such as food supply and weather conditions rather than parental strategy, this interpretation should be treated with caution.